principales sÍndromes neurolÓgicos: … · dose en volumen según transcurrían los años en ......

PRINCIPALES SÍNDROMES NEUROLÓGICOS: EPILEPSIAS

Profesor: Dr. Miguel Ángel Villa Rodríguez FACULTAD DE ESTUDIOS SUPERIORES ZARAGOZA PROGRAMA DE MAESTRÍA Y DOCTORADO EN PSICOLOGÍA

RESIDENCIA EN NEUROPSICOLOGÍA CLÍNICA

[email protected]

INSTITUTO MEXICANO DE LA AUDICIÓN Y EL LENGUAJE

MAESTRÍA EN PATOLOGÍA DE LA AUDICIÓN Y EL

LENGUAJE

1

þ  La epilepsia es una alteración biomédica que produce estallidos anormales de excitación eléctrica en grupos de neuronas que eventualmente puede extenderse a todo el cerebro.

2

þ Esta actividad eléctrica anormal puede tener efectos significativos sobre el funcionamiento cognoscitivo y conductual de las personas que la padecen.

3

S8 www.neurologia.com Rev Neurol 2012; 54 (Supl 3): S7-S18

C. Casas-Fernández

bleciendo una neta diferenciación entre las crisis epilépticas generalizadas y parciales, posteriormen-te denominadas focales, constituyendo, en definiti-va, el andamiaje de las posteriores clasificaciones de las crisis epilépticas y de las epilepsias y síndro-mes epilépticos.

En 1981, en la reunión de la Liga Internacional contra la Epilepsia (ILAE) [3], celebrada en Kioto, nació la primera clasificación de las crisis epilépti-cas que podemos considerar fruto de un consenso internacional, basándose en la conjunción de crite-rios clínicos, electroencefalográficos y del sustrato anatómico (Tabla I), definiendo junto con el con-cepto de crisis generalizadas y parciales las deno-minadas crisis no clasificables. Destaca la subdivi-sión de las crisis parciales en tres fenotipos, inclu-

yendo: simples (sin pérdida de conciencia), comple-jas (con pérdida de conciencia) y secundariamente generalizadas. Estos conceptos se han empleado hasta ahora y han servido para unificar criterios y poder comparar objetivamente las diferentes casuís-ticas, habiendo mostrado, asimismo, una significa-tiva utilidad en el enfoque terapéutico de las epilep-sias en función del fenotipo electroclínico.

En 1989, en la reunión de Nueva Delhi, la ILAE emitió una nueva clasificación [4], dirigida específi-camente a las epilepsias y síndromes epilépticos (Tabla II), estableciendo su etiología en tres catego-rías distintivas: idiopáticas, sintomáticas y criptogé-nicas. Las primeras, idiopáticas, querían traducir un origen genético o posiblemente genético, si bien el desarrollo de la genética y su aplicación en la epilep-sia se encontraba en aquellos años en sus albores, por lo que, obviamente, era un grupo con una defi-nición muy imprecisa. Las sintomáticas reflejaban la existencia de una epilepsia secundaria a una causa demostrable, siendo un grupo que fue incrementán-dose en volumen según transcurrían los años en porcentajes directamente proporcionales al espec-tacular desarrollo de las diferentes exploraciones diagnósticas, y muy destacadamente de las pruebas de neuroimagen. Finalmente, las criptogénicas, de-nominación que etimológicamente procede del grie-go (κρυπτιός, oculto), engloban aquellas epilepsias que se consideraban de origen sintomático, pero en las que no era posible identificar la causa.

Con estas herramientas, el diagnóstico, trata-miento y control evolutivo de las epilepsias se ha sustentado hasta nuestros días, si bien la comunidad epileptológica internacional no ha cesado en reali-zar una crítica constructiva de estas clasificaciones [5-12], buscando una actualización más acorde con los mayores conocimientos etiopatogénicos.

Nuevos intentos semiológicos de clasificar las crisis epilépticas. Clasificación semiológica de Lüders

La clasificación de la ILAE del año 1981 [3] puso las bases elementales para denominar las diferentes posibilidades de manifestación de las crisis epilép-ticas, aunque desde su nacimiento no dejó de con-siderarse la posible modificación de estos concep-tos, siendo 17 años más tarde cuando Lüders et al [6] diseñaron una nueva clasificación semiológica de las crisis (Tabla III), donde éstas son descritas independientemente de los hallazgos electroence-falográficos y de la patología subyacente, pero ofre-ciendo la posibilidad de describir su evolución. En

Tabla I. Clasificación clásica de las crisis epilépticas [3].

Crisis parciales (focales, locales)

Crisis parciales simples (sin afectación de la conciencia)

Con signos motores

Con síntomas somatosensoriales

Con síntomas o signos autonómicos

Con síntomas psíquicos

Crisis parciales complejas (con afectación de la conciencia)

Comienzo parcial simple seguido de afectación de la conciencia

Con afectación de la conciencia desde el comienzo

Crisis parciales con evolución a crisis secundariamente generalizadas

Crisis parciales simples secundariamente generalizadas

Crisis parciales complejas secundariamente generalizadas

Crisis parciales simples que evolucionan a complejas y a secundariamente generalizadas

Crisis generalizadas (convulsivas o no convulsivas)

Ausencias

Ausencias típicas

Ausencias atípicas

Crisis mioclónicas simples o múltiples

Crisis clónicas

Crisis tonicoclónicas

Crisis atónicas (astáticas)

Crisis epilépticas inclasificables

4


C. Casas-Fernández

bleciendo una neta diferenciación entre las crisis epilépticas generalizadas y parciales, posteriormen-te denominadas focales, constituyendo, en definiti-va, el andamiaje de las posteriores clasificaciones de las crisis epilépticas y de las epilepsias y síndro-mes epilépticos.

En 1981, en la reunión de la Liga Internacional contra la Epilepsia (ILAE) [3], celebrada en Kioto, nació la primera clasificación de las crisis epilépti-cas que podemos considerar fruto de un consenso internacional, basándose en la conjunción de crite-rios clínicos, electroencefalográficos y del sustrato anatómico (Tabla I), definiendo junto con el con-cepto de crisis generalizadas y parciales las deno-minadas crisis no clasificables. Destaca la subdivi-sión de las crisis parciales en tres fenotipos, inclu-

yendo: simples (sin pérdida de conciencia), comple-jas (con pérdida de conciencia) y secundariamente generalizadas. Estos conceptos se han empleado hasta ahora y han servido para unificar criterios y poder comparar objetivamente las diferentes casuís-ticas, habiendo mostrado, asimismo, una significa-tiva utilidad en el enfoque terapéutico de las epilep-sias en función del fenotipo electroclínico.

En 1989, en la reunión de Nueva Delhi, la ILAE emitió una nueva clasificación [4], dirigida específi-camente a las epilepsias y síndromes epilépticos (Tabla II), estableciendo su etiología en tres catego-rías distintivas: idiopáticas, sintomáticas y criptogé-nicas. Las primeras, idiopáticas, querían traducir un origen genético o posiblemente genético, si bien el desarrollo de la genética y su aplicación en la epilep-sia se encontraba en aquellos años en sus albores, por lo que, obviamente, era un grupo con una defi-nición muy imprecisa. Las sintomáticas reflejaban la existencia de una epilepsia secundaria a una causa demostrable, siendo un grupo que fue incrementán-dose en volumen según transcurrían los años en porcentajes directamente proporcionales al espec-tacular desarrollo de las diferentes exploraciones diagnósticas, y muy destacadamente de las pruebas de neuroimagen. Finalmente, las criptogénicas, de-nominación que etimológicamente procede del grie-go (κρυπτιός, oculto), engloban aquellas epilepsias que se consideraban de origen sintomático, pero en las que no era posible identificar la causa.

Con estas herramientas, el diagnóstico, trata-miento y control evolutivo de las epilepsias se ha sustentado hasta nuestros días, si bien la comunidad epileptológica internacional no ha cesado en reali-zar una crítica constructiva de estas clasificaciones [5-12], buscando una actualización más acorde con los mayores conocimientos etiopatogénicos.

Nuevos intentos semiológicos de clasificar las crisis epilépticas. Clasificación semiológica de Lüders

La clasificación de la ILAE del año 1981 [3] puso las bases elementales para denominar las diferentes posibilidades de manifestación de las crisis epilép-ticas, aunque desde su nacimiento no dejó de con-siderarse la posible modificación de estos concep-tos, siendo 17 años más tarde cuando Lüders et al [6] diseñaron una nueva clasificación semiológica de las crisis (Tabla III), donde éstas son descritas independientemente de los hallazgos electroence-falográficos y de la patología subyacente, pero ofre-ciendo la posibilidad de describir su evolución. En

Tabla I. Clasificación clásica de las crisis epilépticas [3].

Crisis parciales (focales, locales)

Crisis parciales simples (sin afectación de la conciencia)

Con signos motores

Con síntomas somatosensoriales

Con síntomas o signos autonómicos

Con síntomas psíquicos

Crisis parciales complejas (con afectación de la conciencia)

Comienzo parcial simple seguido de afectación de la conciencia

Con afectación de la conciencia desde el comienzo

Crisis parciales con evolución a crisis secundariamente generalizadas

Crisis parciales simples secundariamente generalizadas

Crisis parciales complejas secundariamente generalizadas

Crisis parciales simples que evolucionan a complejas y a secundariamente generalizadas

Crisis generalizadas (convulsivas o no convulsivas)

Ausencias

Ausencias típicas

Ausencias atípicas

Crisis mioclónicas simples o múltiples

Crisis clónicas

Crisis tonicoclónicas

Crisis atónicas (astáticas)

Crisis epilépticas inclasificables

5

6

225

ESQUEMA DIAGNÓSTICOPROPUESTOPARA LAS PERSONASCON CRISIS EPILÉPTICASY CON EPILEPSIA

Este esquema diagnóstico se divide en5 partes o ejes:

− Eje 1: Consiste en la descripción de lasemiología ictal (durante la crisis).

− Eje 2: Se relaciona con los tipos de crisisque aparecen en la lista de crisisepilépticas y estímulos precipitantes paralas crisis reflejas (ver inciso B).

− Eje 3: Diagnóstico sindrómico que sederiva de la lista de síndromes deepilepsia (ver inciso C).

− Eje 4: Se relaciona con el origen cuandoésta se identifica.

− Eje 5: Se relaciona con el grado deafectación de la función cerebraloriginado por una condición epiléptica(opcional).

Antes de exponer la lista de las defini-ciones y los ejemplos de clasificaciones decrisis epilépticas y síndromes de epilepsiapropuestas por el Grupo de Trabajo de LaLiga Internacional contra la Epilepsia4

realizaremos algunas aclaraciones quepudieran contribuir a una mejor com-prensión de ellas.

Se propone que el término "parcial " sesustituya por el término "focal". El Grupode Trabajo aclara que el término "focal"necesariamente no significa que la regiónepileptogénica sea pequeña y queconstituya un foco bien delimitado deafección neuronal. Las crisis focales y lossíndromes focales se relacionan casisiempre con áreas difusas de disfuncióncerebral.

Otro cambio en la terminología que sepropone es que deben omitirse las palabras"convulsión" y "convulsivo".

También se sugiere que el término"convulsiones febriles" debe ser reem-plazado por "crisis febriles".

El término "espasmo infantil" debe sus-tituirse por el término "espasmos epi-lépticos".

A través de los años el empleo deltérmino "criptogénico" ha sido polémico.Se propuso que aunque continúa siendoaceptable debe ser reemplazado por untérmino más preciso: "probablementesintomático".

A) DEFINICIONES DE TÉRMINOS CLAVE

− Tipo de crisis epiléptica: Evento ictal quese considera representa un mecanismofisiopatológico y un sustrato anatómicoúnicos. Es una entidad diagnóstica conimplicaciones etiológicas, terapéuticas ypronósticas (nuevo concepto).

− Síndrome de epilepsia: Complejo desíntomas y signos que definen una únicacondición de epilepsia. Este debeinvolucrar más que el tipo de crisis: eneste caso, las crisis del lóbulo frontal,por ejemplo, no constituyen un síndrome(concepto cambiado).

− Enfermedad epiléptica: Condición anor-mal con una causa simple y específicabien definida. En este caso la epilepsiamioclónica progresiva es un síndrome,para Unverricht-Lundborg es unaenfermedad (nuevo concepto).

− Encefalopatía epiléptica: Condición en laque las anomalías epilépticas por ellasmismas, se presume que contribuyan aldisturbio progresivo de la funcióncerebral (nuevo concepto).

− Síndrome de epilepsia benigna: Síndromecaracterizado por crisis epilépticas queson tratadas fácilmente, o que norequieren tratamiento y remiten sinsecuela (concepto aclarado).

Capítulo Mexicano

www.epilepsiahoy.com

7

226

− Síndrome de epilepsia refleja: Síndromeen el que todas las crisis epilépticas sonprecipitadas por estímulos sensoriales.Las crisis reflejas son precipitadas porestímulos sensoriales. Las crisis reflejasque ocurren en los síndromes deepilepsias focales y generalizados quetambién se asocian con crisis espon-táneas se consideran como tipos decrisis epilépticas. Las crisis reflejasaisladas también pueden ocurrir ensituaciones que no necesariamenterequieren el diagnóstico de epilepsia. Lascrisis precipitadas por otras circuns-tancias especiales como hipertermia oabstinencia de alcohol no son crisisreflejas (concepto cambiado).

− Crisis focales y síndromes: Reemplaza eltérmino crisis parciales y síndromesrelacionados con localización (términomodificado).

− Crisis epilépticas parciales simples ycomplejas: Estos términos ya no sonrecomendados. La alteración ictal de laconciencia se describirá cuando seaapropiado en crisis individuales, pero noserá empleado para clasificar tipos decrisis específicas (nuevo concepto).

− Síndrome de epilepsia idiopática:Síndrome que es solamente epilepsia, sinlesión estructural de base cerebral u otrossíntomas o signos neurológicos. Sepresume que son genéticos ydependientes de la edad (término nomodificado).

− Síndrome de epilepsia sintomática:Síndrome en el que las crisis epilépticasson el resultado de una o más lesiónestructural identificables del cerebro(término no modificado).

− Síndrome de epilepsia probablementesintomática: Sinónimo con el términocriptogénico, pero se prefiere éste,empleado para definir síndromes que sepresumen sean sintomáticos, pero cuyoorigen no ha sido identificado (nuevotérmino).

B) TIPOS DE CRISIS EPILÉPTICASY CRISIS PRECIPITADASPOR ESTÍMULOS REFLEJOS

1. Crisis autolimitadas:

− Generalizadas:• Tónico-clónicas (incluye variacio-

nes que comienzan con una faseclónica o mioclónica).

• Clónicas: sin características tónicas;con características tónicas.

• Ausencias típicas.• Ausencias atípicas.• Ausencias mioclónicas.• Tónicas.• Espasmos.• Mioclónicas.• Mioclonías del globo ocular: sin

ausencias; con ausencias• Miotónicas atónicas.• Mioclono negativo.• Atónicas.• Reflejas en síndromes epilépticos

generalizados− Focales:• Sensoriales focales:

Con síntomas sensoriales elementa-les (Ej.: crisis del lóbulo parietal yoccipital).Con síntomas experienciales (Ej.:crisis témporo-parieto-occipitales).• Motoras focales:

Con signos motores elementales.Con signos motores tónicos asimé-tricos (Ej.: crisis motoras suplemen-tarias).Con automatismos (del lóbulotemporal) típicos (Ej.: crisis dellóbulo temporal mesial).Con automatismos hipercinéticos.Con mioclono negativo focal.Con crisis motoras inhibitorias.Gelásticas.Hemiclónicas.

226

− Síndrome de epilepsia refleja: Síndromeen el que todas las crisis epilépticas sonprecipitadas por estímulos sensoriales.Las crisis reflejas son precipitadas porestímulos sensoriales. Las crisis reflejasque ocurren en los síndromes deepilepsias focales y generalizados quetambién se asocian con crisis espon-táneas se consideran como tipos decrisis epilépticas. Las crisis reflejasaisladas también pueden ocurrir ensituaciones que no necesariamenterequieren el diagnóstico de epilepsia. Lascrisis precipitadas por otras circuns-tancias especiales como hipertermia oabstinencia de alcohol no son crisisreflejas (concepto cambiado).

− Crisis focales y síndromes: Reemplaza eltérmino crisis parciales y síndromesrelacionados con localización (términomodificado).

− Crisis epilépticas parciales simples ycomplejas: Estos términos ya no sonrecomendados. La alteración ictal de laconciencia se describirá cuando seaapropiado en crisis individuales, pero noserá empleado para clasificar tipos decrisis específicas (nuevo concepto).

− Síndrome de epilepsia idiopática:Síndrome que es solamente epilepsia, sinlesión estructural de base cerebral u otrossíntomas o signos neurológicos. Sepresume que son genéticos ydependientes de la edad (término nomodificado).

− Síndrome de epilepsia sintomática:Síndrome en el que las crisis epilépticasson el resultado de una o más lesiónestructural identificables del cerebro(término no modificado).

− Síndrome de epilepsia probablementesintomática: Sinónimo con el términocriptogénico, pero se prefiere éste,empleado para definir síndromes que sepresumen sean sintomáticos, pero cuyoorigen no ha sido identificado (nuevotérmino).

B) TIPOS DE CRISIS EPILÉPTICASY CRISIS PRECIPITADASPOR ESTÍMULOS REFLEJOS

1. Crisis autolimitadas:

− Generalizadas:• Tónico-clónicas (incluye variacio-

nes que comienzan con una faseclónica o mioclónica).

• Clónicas: sin características tónicas;con características tónicas.

• Ausencias típicas.• Ausencias atípicas.• Ausencias mioclónicas.• Tónicas.• Espasmos.• Mioclónicas.• Mioclonías del globo ocular: sin

ausencias; con ausencias• Miotónicas atónicas.• Mioclono negativo.• Atónicas.• Reflejas en síndromes epilépticos

generalizados− Focales:• Sensoriales focales:

Con síntomas sensoriales elementa-les (Ej.: crisis del lóbulo parietal yoccipital).Con síntomas experienciales (Ej.:crisis témporo-parieto-occipitales).• Motoras focales:

Con signos motores elementales.Con signos motores tónicos asimé-tricos (Ej.: crisis motoras suplemen-tarias).Con automatismos (del lóbulotemporal) típicos (Ej.: crisis dellóbulo temporal mesial).Con automatismos hipercinéticos.Con mioclono negativo focal.Con crisis motoras inhibitorias.Gelásticas.Hemiclónicas.

8

227

Generalizadas secundariamente.Reflejas en síndromes epilépticosfocales.

Crisis continuas:− Estado de mal epiléptico generalizado:• Tónico-clónico generalizado• Clónica.• De ausencia.• Tónico.• Mioclónico.

− Estado de mal epiléptico focal:• Epilepsia partialis continua de

Kojevnikov.• Aura continua.• Límbico (psicomotor).• Hemiconvulsivo con hemiparesia.

Crisis reflejas y estímulos precipitantes:− Visuales:• Luminosos: el color debe ser espe-

cificado cuando sea posible.• A patrón.• Otros.

− Pensamiento:• Música.• Comer.• Praxia.• Somatosensorial.• Propioceptivo.• Lectura.• Agua caliente.• Sobresalto.

C) SÍNDROME DE EPILEPSIA

Epilepsias focales:− Epilepsias idiopáticas de la lactante y

de la infancia:• Crisis de la lactancia benignas (no

familiares).• Epilepsia benigna de la infancia con

puntas centrotemporales.

• Epilepsia occipital benigna de lainfancia de comienzo precoz (tipoPanayilotopoulos).

• Epilepsia occipital de la infancia decomienzo tardío (tipo Gastaut).

− Epilepsias familiares (autosómicasdominantes):• Crisis neonatales familiares benig-

nas.• Crisis de la lactancia familiares be-

nignas.• Epilepsia del lóbulo frontal noctur-

na autosómica dominante.• Epilepsia del lóbulo temporal fami-

liar.• Epilepsia familiar con focos varia-

bles (en desarrollo).− Epilepsias sintomáticas (o probable-

mente sintomáticas):• Epilepsias límbicas• Epilepsia del lóbulo temporal mesial

con esclerosis hipocámpica• Epilepsia del lóbulo temporal mesial

definida por etiologías definidas

Otros tipos definidos por localización yetiologías− Epilepsias neocorticales:• Síndrome de Rasmussen.• Síndrome hemiplejia-hemiconvul-

sión.• Otros tipos definidos por localiza-

ción y etiologías.• Crisis parciales migratorias de la

lactancia temprana (en desarrollo).− Epilepsias generalizadas idiopáticas:• Epilepsia mioclónica benigna en la

lactancia.• Epilepsia con crisis mioclónicas-as-

tásicas.• Epilepsia ausencia de la infancia.• Epilepsia con ausencias mioclónicas.• Epilepsias generalizadas con feno-

tipos variables:

227

Generalizadas secundariamente.Reflejas en síndromes epilépticosfocales.

Crisis continuas:− Estado de mal epiléptico generalizado:• Tónico-clónico generalizado• Clónica.• De ausencia.• Tónico.• Mioclónico.

− Estado de mal epiléptico focal:• Epilepsia partialis continua de

Kojevnikov.• Aura continua.• Límbico (psicomotor).• Hemiconvulsivo con hemiparesia.

Crisis reflejas y estímulos precipitantes:− Visuales:• Luminosos: el color debe ser espe-

cificado cuando sea posible.• A patrón.• Otros.

− Pensamiento:• Música.• Comer.• Praxia.• Somatosensorial.• Propioceptivo.• Lectura.• Agua caliente.• Sobresalto.

C) SÍNDROME DE EPILEPSIA

Epilepsias focales:− Epilepsias idiopáticas de la lactante y

de la infancia:• Crisis de la lactancia benignas (no

familiares).• Epilepsia benigna de la infancia con

puntas centrotemporales.

• Epilepsia occipital benigna de lainfancia de comienzo precoz (tipoPanayilotopoulos).

• Epilepsia occipital de la infancia decomienzo tardío (tipo Gastaut).

− Epilepsias familiares (autosómicasdominantes):• Crisis neonatales familiares benig-

nas.• Crisis de la lactancia familiares be-

nignas.• Epilepsia del lóbulo frontal noctur-

na autosómica dominante.• Epilepsia del lóbulo temporal fami-

liar.• Epilepsia familiar con focos varia-

bles (en desarrollo).− Epilepsias sintomáticas (o probable-

mente sintomáticas):• Epilepsias límbicas• Epilepsia del lóbulo temporal mesial

con esclerosis hipocámpica• Epilepsia del lóbulo temporal mesial

definida por etiologías definidas

Otros tipos definidos por localización yetiologías− Epilepsias neocorticales:• Síndrome de Rasmussen.• Síndrome hemiplejia-hemiconvul-

sión.• Otros tipos definidos por localiza-

ción y etiologías.• Crisis parciales migratorias de la

lactancia temprana (en desarrollo).− Epilepsias generalizadas idiopáticas:• Epilepsia mioclónica benigna en la

lactancia.• Epilepsia con crisis mioclónicas-as-

tásicas.• Epilepsia ausencia de la infancia.• Epilepsia con ausencias mioclónicas.• Epilepsias generalizadas con feno-

tipos variables:

9

•  Epilepsia del lóbulo frontal

•  Epilepsia del lóbulo temporal

•  Epilepsia del lóbulo parietal

•  Epilepsia del lóbulo occipital

10

Factores que determinan las alteraciones cognoscitivas

Psicológicos

Debidos al tratamiento

Biológicos

FACTORES DE LAS ALTERACIONES COGNOSCITIVAS

1.  Etiología 2.  Tipo de crisis 3.  Edad de inicio 4.  Frecuencia 5.  Duración y severidad 6.  Alteraciones funcionales fisiológicas intra e interictales

producidas por las crisis.

11

FACTORES DE LAS ALTERACIONES COGNOSCITIVAS

7.  Daño cerebral estructural producido por la recurrencia de las crisis.

8.  Factores hereditarios 9.  Factores psicosociales 10.  Efecto de los medicamentos

12

þ Aun cuando la mayoría de las personas con epilepsia tienen una inteligencia normal, como grupo son más susceptibles de sufrir alteraciones en algunos procesos cognoscitivos en comparación con personas de su misma edad y escolaridad que no tienen epilepsia.

þ Las alteraciones cognoscitivas dependen de la neuropatología subyacente

13

þ En la epilepsia del lóbulo temporal pueden presentarse problemas de memoria y en las epilepsias focales que afectan la zona del lenguaje pueden alterarse la capacidad para encontrar las palabras o para denominar.

þ Pero además de los efectos adversos de las crisis en sí mismas, hay una serie de factores que determinan las alteraciones cognoscitivas.

14

La queja más frecuente de las personas con epilepsia es que tienen problemas de memoria

•  ¿Por qué puedo acordarme de cosas que ocurrieron hace mucho tiempo, pero no puedo retener lo que la gente me dice o recordar lo que pasó ayer?

•  ¿Cómo puede ser que reconozca las caras de las personas sin ningún problema y, sin embargo, no recuerde sus nombres, aunque las conozca desde hace mucho tiempo?

•  ¿Por qué, aunque conozca una determinada información perfectamente, mi mente se queda en blanco cuando necesito emplearla?

15

ALTERACIONES DE LA MEMORIA Y DEL APRENDIZAJE EN LA EPILEPSIA DEL LÓBULO TEMPORAL

16

Cognitive phenotypes in temporal lobe epilepsy

BRUCE HERMANN,1 MICHAEL SEIDENBERG,2 EUN-JEONG LEE,3 FONG CHAN,3

and PAUL RUTECKI1

1Department of Neurology, University of Wisconsin-Madison, Madison, Wisconsin2Department of Psychology, Rosalind Franklin University of Medicine and Science, North Chicago, Illinois3Rehabilitation Psychology, University of Wisconsin-Madison, Madison, Wisconsin

(Received Febrary 22, 2006; Final Revision July 17, 2006; Accepted July 18, 2006)

Abstract

The objective of this study is to determine if distinct cognitive phenotypes could be identified in temporal lobeepilepsy. Epilepsy patients (n5 96) and healthy controls (n5 82) underwent comprehensive neuropsychologicalassessment. Adjusted (age, gender, and education) test scores for epilepsy subjects were grouped into cognitivedomains (intelligence, language, visuoperception, immediate and delayed memory, executive function, andcognitive0psychomotor speed). Cluster analysis revealed three distinct cognitive profiles types: (1) minimallyimpaired (47% of subjects); (2) memory impaired (24%); and (3) memory, executive, and speed impaired (29%).The three cluster groups exhibited different patterns of results on demographic, clinical epilepsy, brain volumetrics,and cognitive course over a 4-year interval. The specific profile characteristics of the identified cognitivephenotypes are presented and their implications for the investigation of the neurobehavioral complications ofepilepsy are discussed. (JINS, 2007, 13, 12–20.)

Keywords: Cluster analysis, Neuropsychological tests, MRI, Progression, Memory, Seizures

INTRODUCTION

Neuropsychological impairment is an important co-morbidityof chronic epilepsy (Elger et al., 2004). Considerableresearch has examined the relationship between cognitionand a variety of clinical factors including etiology, age ofonset, seizure type and severity, duration, antiepilepsy med-ications, and other factors. (Aldenkamp & Arends, 2004;Dodrill, 2004; Helmstaedter & Kurthen, 2001; Jones-Gotman, 2000; Saling et al., 1993). In addition, modal cog-nitive profiles have been derived for various epilepsysyndromes (including mesial temporal lobe epilepsy), andefforts have been undertaken to identify the shared versusunique cognitive risks across epilepsy syndromes (Elgeret al., 2004; Lassonde et al., 2000; Nolan et al., 2003).These approaches have provided insight into the influenceof clinical seizure factors on cognition in epilepsy.

A yet untapped approach to understanding cognitive mor-bidity in epilepsy is taxonomic in nature. This involvesaddressing the question of whether empirically derived

groupings of patients with similar profiles of cognitivefunction can be identified either within or across epilepsysyndromes. Taxonomies facilitate reliable clustering of indi-viduals into meaningful groups, provide a common lan-guage and organizing influence in the field, and they set thestage for further investigation of clinical and neurobiolog-ical correlates. Significant progress in the epilepsies hasresulted from efforts to identify and characterize the hetero-geneity inherent in the disorder, perhaps the best examplebeing the classification of seizures and epilepsy syndromes(Commission on Classification and Terminology of theILAE, 1981, 1989; Duchowny & Harvey, 1996; Engel, 2001;Fisher et al., 2005; Luders et al., 1998).

To date, taxonomic approaches have rarely been used toadvance understanding of the neurobehavioral complica-tions of the epilepsies (Paradiso et al., 1994). That is, ratherthan grouping patients based on clinical seizure character-istics (e.g., seizure frequency) and examining the relation-ships of individual clinical seizure characteristics tocognition, one derives a grouping of patients based solelyon their pattern of performance across several cognitivedomains. Such an approach would identify distinct cogni-tive profile types, the relative proportion of patients express-ing each profile type, and provide a basis for identifying the

Correspondence and reprint requests to: Dr. Bruce Hermann, Depart-ment of Neurology, University of Wisconsin-Madison, 600 N. Highland,Madison, Wisconsin 53792. E-mail: [email protected]

Journal of the International Neuropsychological Society (2007), 13, 12–20.Copyright © 2007 INS. Published by Cambridge University Press. Printed in the USA.DOI: 10.10170S135561770707004X

12

17 Cognitive phenotypes in temporal lobe epilepsy

BRUCE HERMANN,1 MICHAEL SEIDENBERG,2 EUN-JEONG LEE,3 FONG CHAN,3

and PAUL RUTECKI1

1Department of Neurology, University of Wisconsin-Madison, Madison, Wisconsin2Department of Psychology, Rosalind Franklin University of Medicine and Science, North Chicago, Illinois3Rehabilitation Psychology, University of Wisconsin-Madison, Madison, Wisconsin

(Received Febrary 22, 2006; Final Revision July 17, 2006; Accepted July 18, 2006)

Abstract

The objective of this study is to determine if distinct cognitive phenotypes could be identified in temporal lobeepilepsy. Epilepsy patients (n5 96) and healthy controls (n5 82) underwent comprehensive neuropsychologicalassessment. Adjusted (age, gender, and education) test scores for epilepsy subjects were grouped into cognitivedomains (intelligence, language, visuoperception, immediate and delayed memory, executive function, andcognitive0psychomotor speed). Cluster analysis revealed three distinct cognitive profiles types: (1) minimallyimpaired (47% of subjects); (2) memory impaired (24%); and (3) memory, executive, and speed impaired (29%).The three cluster groups exhibited different patterns of results on demographic, clinical epilepsy, brain volumetrics,and cognitive course over a 4-year interval. The specific profile characteristics of the identified cognitivephenotypes are presented and their implications for the investigation of the neurobehavioral complications ofepilepsy are discussed. (JINS, 2007, 13, 12–20.)

Keywords: Cluster analysis, Neuropsychological tests, MRI, Progression, Memory, Seizures

INTRODUCTION

Neuropsychological impairment is an important co-morbidityof chronic epilepsy (Elger et al., 2004). Considerableresearch has examined the relationship between cognitionand a variety of clinical factors including etiology, age ofonset, seizure type and severity, duration, antiepilepsy med-ications, and other factors. (Aldenkamp & Arends, 2004;Dodrill, 2004; Helmstaedter & Kurthen, 2001; Jones-Gotman, 2000; Saling et al., 1993). In addition, modal cog-nitive profiles have been derived for various epilepsysyndromes (including mesial temporal lobe epilepsy), andefforts have been undertaken to identify the shared versusunique cognitive risks across epilepsy syndromes (Elgeret al., 2004; Lassonde et al., 2000; Nolan et al., 2003).These approaches have provided insight into the influenceof clinical seizure factors on cognition in epilepsy.

A yet untapped approach to understanding cognitive mor-bidity in epilepsy is taxonomic in nature. This involvesaddressing the question of whether empirically derived

groupings of patients with similar profiles of cognitivefunction can be identified either within or across epilepsysyndromes. Taxonomies facilitate reliable clustering of indi-viduals into meaningful groups, provide a common lan-guage and organizing influence in the field, and they set thestage for further investigation of clinical and neurobiolog-ical correlates. Significant progress in the epilepsies hasresulted from efforts to identify and characterize the hetero-geneity inherent in the disorder, perhaps the best examplebeing the classification of seizures and epilepsy syndromes(Commission on Classification and Terminology of theILAE, 1981, 1989; Duchowny & Harvey, 1996; Engel, 2001;Fisher et al., 2005; Luders et al., 1998).

To date, taxonomic approaches have rarely been used toadvance understanding of the neurobehavioral complica-tions of the epilepsies (Paradiso et al., 1994). That is, ratherthan grouping patients based on clinical seizure character-istics (e.g., seizure frequency) and examining the relation-ships of individual clinical seizure characteristics tocognition, one derives a grouping of patients based solelyon their pattern of performance across several cognitivedomains. Such an approach would identify distinct cogni-tive profile types, the relative proportion of patients express-ing each profile type, and provide a basis for identifying the

Correspondence and reprint requests to: Dr. Bruce Hermann, Depart-ment of Neurology, University of Wisconsin-Madison, 600 N. Highland,Madison, Wisconsin 53792. E-mail: [email protected]

Journal of the International Neuropsychological Society (2007), 13, 12–20.Copyright © 2007 INS. Published by Cambridge University Press. Printed in the USA.DOI: 10.10170S135561770707004X

12

one stage to the next at the three stages preceding the four-cluster stage, compared with increases of 110, 118, and 678at the three subsequent stages in the hierarchy. Thus, clusterhomogeneity dropped substantially after the four-clusterstage, however, two of the clusters in the four-cluster solu-tion had very small sample sizes (n 5 11 and n 5 16) and

differed primarily in the severity of abnormality in execu-tive and motor domains. In the three cluster solution, thesetwo groups were merged to form a larger cluster. The three-cluster solution provided a reasonable compromise betweenmaximizing cluster homogeneity, deriving a cluster solu-tion resulting in groups of epilepsy patients that could bemeaningfully interpreted and with appropriate sample sizesfor subsequent statistical analyses.

RESULTS

The mean cognitive performance for the three cluster groupsis provided in Table 3. A graphic representation of the cog-nitive profiles is presented in Figure 1. Mean performancefor the controls (mean 5 0 and SD 5 1) is represented onthe x-axis. A 1-way MANOVA was used to compare thethree cluster groups across the seven cognitive domains. Asignificant effect of group was obtained, Hotelling T54.13,df 5 21,491, p, .001, and univariate effects were signifi-cant across all cognitive domains. A description of the cog-nitive profile associated with each of the cluster groups isprovided later.

Cluster 1: minimally impaired group

Cluster 1 was composed of 44 participants (47% of tempo-ral lobe group) who exhibited a pattern of minimal cogni-tive impairment compared to controls. There were nosignificant differences in intelligence, perception, and imme-diate memory. However, Cluster 1 exhibited statisticallysignificant lower scores than controls in the domains oflanguage ( p5 .002), delayed memory ( p5 .01), executivefunction ( p5 .025) and cognitive0psychomotor speed ( p5.006). Within group pairwise contrasts indicated that thedelayed memory ( p5 .011), language ( p5 .001) and exec-utive function ( p5 .006) domains were significantly lowerthan mean IQ.

Table 3. Means and standard deviations of the cognitivemeasures for the clusters

Cognitive measure Cluster MeanStandarddeviation

Percentimpaired

at z2 1.64

Intelligence 1 20.19 0.90 2.32 21.14 0.48 8.73 22.00 0.83 63.0

Language 1 20.56 0.89 11.42 20.75 0.87 13.03 22.32 1.50 59.3

Visuoperception 1 20.08 0.57 0.02 21.05 0.83 13.03 21.74 1.40 55.6

Memory-Immediate 1 20.31 0.68 6.82 22.14 0.81 69.63 23.13 0.89 92.6

Memory-Delayed 1 20.38 0.65 4.52 21.99 0.98 69.63 22.71 0.86 92.6

Executive 1 20.44 0.94 13.62 20.52 0.8 34.33 23.18 1.97 77.8

Motor 1 20.75 0.94 22.72 20.74 1.03 17.43 24.36 1.86 85.2

Fig. 1. Mean cluster performance across cognitive domains.

Cognitive phenotypes in temporal lobe epilepsy 15

18

Cluster 1: Alteraciones mínimas Cluster 2: Alteraciones de memoria Cluster 3: Alteraciones de memoria, de FE y de velocidad de procesamiento

Cluster 2: memory impaired group

Cluster 2 was composed of 23 participants (24% of tempo-ral lobe group) who exhibited prominent memory impair-ment (approximately two standard deviations below theaverage scores of the controls) in the context of mild-moderate cognitive impairment in the remaining domains.Although the Cluster 2 group obtained significantly lowerscores on intelligence ( p, .001) and visuoperception ( p,.001) than the Cluster 1 group, memory (both immediateand delayed) was most significantly affected. Within grouppairwise contrasts confirmed that only immediate ( p5 .034)and delayed ( p 5 .023) memory were significantly lowerthan mean IQ. The memory domains were also signifi-cantly lower compared to all other cognitive domains. Fur-thermore, Cluster 2 subjects scored similar to Cluster 1subjects on the executive and speed domains ( p’s. .05).

Cluster 3: memory, executive, andspeed impaired group

Cluster 3 was composed of 27 participants (29% of tempo-ral lobe group) who exhibited a pattern of moderate to severecognitive impairment. Cluster 3 subjects performed signif-icantly ( p, .001) worse than controls across all cognitivedomains, and also performed significantly worse than bothCluster 1 and 2 groups across all cognitive domains (allp’s, .007). In the context of this generalized impairment,within group pairwise comparisons indicated that com-pared to IQ, Cluster 3 subjects exhibited significant impair-ments in memory ( p, .004), speeded psychomotor ability( p, .001), and executive function ( p, .001).

Cluster Characteristics

Demographic and clinical seizure features

The three cluster groups were compared on demographiccharacteristics (age, gender, and education), clinical epi-lepsy features (age of onset, duration of disorder, number ofantiepilepsy medications), and ICV-adjusted quantitativeMRI variables (total cerebral gray and white matter, totalcerebral CSF, total hippocampal volume). These compari-sons were conducted using one-way analysis of variancewith post-hoc pair-wise comparisons. Table 4 provides groupmeans for the demographic and clinical seizure variables,and Figure 2 provides a depiction of ICV adjusted volumet-ric measurements across groups.

There were no significant differences between the threecluster groups on gender, age of epilepsy onset, education,and overall seizure frequency (daily, weekly, monthly, andyearly). A subset of the study sample underwent ictal mon-itoring, which identified patients with unilateral left (n 524) or right (n5 21) temporal lobe onset. The distribution

Fig. 2. Mean ICV adjusted z-scores of quantitative volumetric measurements across cluster groups.

Table 4. Cluster characteristics

1 2 3

Age 33.6 37.9 41.8Gender (% F) 63 72 65Education 13.4 13.1 12.4Age of onset 15.6 15.6 12.9Duration 17.5 21.7 27.7AEDs 1.6 1.7 2

Note: Age, education, onset, duration and AEDs are presented as means.

16 B. Hermann et al.

19

Quejas de memoria

•  Los déficits de la memoria son la preocupación clínica más importante de los pacientes con epilepsia y de sus familiares.

• Sin embargo rara vez se puede identificar una sola variable que explique los déficits de memoria y del aprendizaje.

20

Déficits de la memoria

y del aprendizaje

Atención

Número de crisis

Edad de inicio de las crisis

Frecuencia y severidad de

estatus epilépticus

previos

Efectos adversos de los medicamentos

Alteraciones psiquiátricas

21

¿Qué es la memoria?

þ La memoria está formada por diferentes dimensiones o sistemas.

þ Puede afectarse alguna de estas funciones y mantenerse otras.

22

DIMENSIONES DE LA MEMORIA

Tiempo

Inmediata, operativa o de

trabajo

A largo plazo

Capacidad para demorar el recuerdo

Memoria reciente

Memoria remota Prospectiva

23


Tipo de información

Conocimientos generales (mem

semántica)

Experiencias personales (mem

episódica o autobiográfica)

Habilidades y procedimientos

24


Formato

Verbal

Visual

Táctil, etc.

25

26

Déficits de

memoria en la ELT

Memoria de

trabajo Memoria episódica

ETAPAS DE LA MEMORIA

Etapas

Codificación

Almacenamiento

Recuperación

27

28

This effect of visual similarity on span for verbal materialspresents something of a problem; given that it occurs under

standard non-suppressed conditions, it indicates that visualand phonological information are combined in some way. The

Review B a d d e l e y – T h e e p i s o d i c b u f f e r i n w o r k i n g m e m o r y

418T r e n d s i n C o g n i t i v e S c i e n c e s – V o l . 4 , N o . 1 1 , N o v e m b e r 2 0 0 0

The term working memory is used in at least three differentways in different areas of cognitive science. It is used here, andin cognitive psychology generally to refer to a limited capacitysystem allowing the temporary storage and manipulation of in-formation necessary for such complex tasks as comprehension,learning and reasoning (Refs a,b). In the animal learning lab-oratory the term refers to the storage of information across sev-eral trials performed within the same day, as demanded by taskssuch as the radial arm maze (Ref. c). In artificial intelligence,production system architectures apply the term to the com-ponent, often unlimited in capacity, that is assumed to be responsible for holding the productions (Ref. d).

These three meanings are thus not interchangeable.Performance of rats on a the radial arm maze, for example,probably relies upon long-term memory (LTM), while theunlimited capacity of the working memory component typi-cally assumed by production system architectures differsmarkedly from the capacity limitation assumed by most of themodels proposed within cognitive psychology.

The multi-component model of working memory (WM) thatforms the basis of this review developed from an earlier conceptof short-term memory (STM), that was assumed to comprise aunitary temporary storage system. This approach was typified bythe model of Atkinson and Shiffrin (Ref. e). However, their

model encountered problems (1) in accounting for the relation-ship between type of encoding and LTM (Ref. f), (2) in explain-ing why patients with grossly defective STM had apparently normal LTM, and (3) in accounting for the effects of a range of con-current tasks on learning, comprehending and reasoning (Ref. b).

Baddeley and Hitch proposed the three-component WMmodel (shown in Fig. Ia) to account for this pattern of data. Themodel comprised an attentional control system, the ‘centralexecutive’, aided by two subsidiary slave systems, the ‘phono-logical loop’ and the ‘visuospatial sketchpad’ (Ref. b). The loopis assumed to hold verbal and acoustic information using a tem-porary store and an articulatory rehearsal system, which clinicallesion studies, and subsequently neuroradiological studies, sug-gested are principally associated with Brodmann areas, 40 and44 respectively. The sketchpad is assumed to hold visuospatialinformation, to be fractionable into separate visual, spatial andpossibly kinaesthetic components, and to be principally repre-sented within the right hemisphere (areas 6, 19, 40 and 47). Thecentral executive is also assumed to be fractionable. Although itis less well understood, frontal lobe areas appear to be stronglyimplicated. An excellent recent overview of short-term andworking memory is given by Gathercole (Ref. g).

Working memory and long-term memory were initiallytreated as quite separate because patients with clear short-termphonological deficits appear to have intact LTM (Ref. b).Subsequent research has shown that such patients do have spe-cific deficits in long-term phonological learning, for example,learning the vocabulary of a new language (Ref. h). Further evi-dence based on the link between phonological loop perfor-mance and vocabulary level in children, suggests that the loopmight have evolved to enhance language acquisition (Ref. h).As predicted by this supposition, patients with phonologicalloop deficits have great difficulty in acquiring novel vocabulary.It seems likely that a similar function is served by the visuospa-tial sketchpad, although there is as yet little investigation of thistopic. If one accepts the hypothesis of an equivalent functionfor the sketchpad, possibly in acquiring visuospatial semantics,then the framework is modified to that shown in Fig. Ib.

References

a Miller, G.A. et al. (1960) Plans and the Structure of Behavior, Holt

b Baddeley, A.D. and Hitch, G. (1974) Working memory. In The

Psychology of Learning and Motivation (Bower, G.A., ed.),

pp. 48–79, Academic Press

c Olton, D.S. et al. (1980) Hippocampal function: working memory

or cognitive mapping? Physiol. Psychol. 8, 239–246

d Newell, A. and Simon, H.A. (1972) Human Problem Solving,

Prentice-Hall

e Atkinson, R.C. and Shiffrin, R.M. (1968) Human memory: a

proposed system and its control processes. In The Psychology of

Learning and Motivation: Advances in Research and Theory

(Spence, K.W., ed.), pp. 89–195, Academic Press

f Craik, F.I.M. and Lockhart, R.S. (1972) Levels of processing: a

framework for memory research. J. Verbal Learn. Verbal Behav.

11, 671–684

g Gathercole, S.E. (1999) Cognitive approaches to the development

of short-term memory. Trends Cognit. Sci. 3, 410–419

h Baddeley, A.D. et al. (1998) When long-term learning depends on

short-term storage. J. Mem. Lang. 27, 586–595

i Cattell, R.B. (1963) Theory of fluid and crystallized intelligence: a

critical experiment. J. Educ. Psychol. 54, 1–22

Box 1. The concept of working memory

Visuospatialsketchpad

Centralexecutive

Phonologicalloop

Centralexecutive

Visuospatialsketchpad

Visualsemantics

EpisodicLTM Language

Phonologicalloop

(a)

(b)

trends in Cognitive SciencesFig. I. (a) The initial three-component model of workingmemory proposed by Baddeley and Hitch (Ref. b). The three-component model assumes an attentional controller, the centralexecutive, aided by two subsidiary systems, the phonological loop,capable of holding speech-based information, and the visuospa-tial sketchpad, which performs a similar function for visual infor-mation. The two subsidiary systems themselves form active storesthat are capable of combining information from sensory input,and from the central executive. Hence a memory trace in thephonological store might stem either from a direct auditory input,or from the subvocal articulation of a visually presented item suchas a letter. (b) A further development of the WM model. Itbecame clear that the phonological loop plays an important rolein long-term phonological learning, in addition to short-term stor-age. As such it is associated with the development of vocabularyin children, and with the speed of acquisition of foreign languagevocabulary in adults. The shaded areas represent ‘crystallized’ cog-nitive systems capable of accumulating long-term knowledge(e.g. language and semantic knowledge). Unshaded systems areassumed to be ‘fluid’ capacities, such as attention and temporarystorage, and are themselves unchanged by learning, other thanindirectly via the crystallized systems (Ref. i).

Baddeley, A. The episodic buffer: A new component of working memory? Trends in Cognitive Sciences – Vol. 4, No. 11, November 2000

Tareas de MT en ELT

•  Tareas motoras visoespaciales

• Span de letras o dígitos

• Cubos de Corsi

•  Igualación a la muestra demorada

•  Laberintos

29

Factores relacionados con la severidad de la alteración de la MT en ELT

1.  Número de crisis / Inicio temprano de la ELT

2.  Lateralidad. ELTI: peor ejecución en MT verbal

3.  Esclerosis del hipocampo

30

31

[28] or WAIS-R [29], Trails A (TrailsA) and Trails B (TrailsB) [30],the Color/Word interference score on the Stroop Task (StroopCW)[31], the number of perseverative errors and number of categoriesachieved on the Wisconsin Card Sort Test (WCSTPer and WCSTCat)[32], and the number of words generated on a verbal fluency test(WordFlu) [33]. The aforementioned tests are commonly used neu-ropsychological measures of working memory and executive func-tioning [34].

3. Results

The resulting sample consisted of 207 individuals with con-firmed TLE and 217 individuals with PNES (see Table 1 for demo-graphic data). Within the TLE sample, 129 individuals wereconfirmed to have MTLE (approximately 62%; the distinction be-tween TLE and MTLE was based on the presence of MRI findingsof hippocampal atrophy and T2 signal change suggestive of medialtemporal sclerosis in patients with MTLE). Seizure calendars wereused to calculate the average monthly seizure frequency. Mean sei-zure frequency (used to calculate lifetime seizure load) for TLE was22 per month (range: <1/month to 750/month); mean seizure fre-quency for PNES was 78/month (range: <1/month to 4875/month).It should be noted that there were five outliers between the twodiagnostic groups. Three individuals in the TLE group were havingan average of 25 simple partial seizures per day, 13 simple partialseizures per day, and 9 complex partial seizures or generalized to-nic–clonic seizures per day, respectively. Within the PNES group,two individuals reported having 162 nonepileptic seizures perday and 120 nonepileptic seizures per day, respectively. These fiveindividuals were included in the analyses because one of the pur-poses of the study was to understand the relationship between sei-zure frequency and cognitive performance. To remove these fiveindividuals from the analyses would reduce our ability to see theeffects of a large number of seizures on cognitive performance.

Data screening was conducted to assess data accuracy, potentialoutliers, and assumptions of normality for both predictor anddependent variables. Univariate normality was assessed withskewness and kurtosis coefficients, Q–Q-Plots, and histograms.Skewness and kurtosis values above +1.00 or below –1.00 wereused as an indication that the values were not normally distributed[35]. Three dependent variables were not normally distributed:TrailsA, TrailsB, and WCSTPer. Each of the three was transformedby taking the natural log. Two predictors, duration of epilepsy/PNES (Duration) and prospective lifetime seizure frequency (LifeF-

req) were not normally distributed. These positively skewed pre-dictors were transformed in the following manner: (1)Duration = ln(Duration + 1) and (2) LifeFreq = ln(LifeFreq1/4).

All correlations among the dependent variables were signifi-cantly different from zero (P < 0.01). Moderate correlations be-tween the dependent variables were expected as the tests are allmeasures of higher cognitive functioning abilities. Higher scoreson the WordFlu, StroopCW, WCSTCat, and deviation quotientswere associated with reduced time to complete TrailsA and TrailsBas well as fewer perseverative errors on the WCST (WCSTPer). Asexpected, those subtests from the same overall test (i.e., TrailsAand TrailsB, WMDQ and VCDQ, and PCDQ) were correlated thehighest. Correlations among the predictors Group, AgeatOnset,Education, Duration, and LifeFreq were all significantly differentfrom zero and ranged from –0.67 to 0.56 with the exception ofEducation and Duration, which were not correlated. As expected,there was a direct relationship between AgeatOnset, Duration,and LifeFreq. Duration decreased as AgeatOnset increased, andthe fewer years an individual had seizures, the lower the LifeFreq.Descriptive statistics for the dependent variables after transforma-tion (means and SD of the neuropsychological tests) are listed inTable 2. The cross-sectional design of the study resulted in onlyone neuropsychological testing time point per patient. Descriptivestatistics for the predictors were: AgeatOnset (M = 24.26,SD = 14.48), Duration in years before transformation (M = 12.82,SD = 12.80) and after (M = 2.11, SD = 1.10) transformation, andLifeFreq before (M = 4564.99, SD = 25249.50) and after (M = 4.82,SD = 2.06) transformation.

There were 86 missing values out of a total of 3870 values(approximately 2%). Fifty patients had one or more missing values.In many cases, seizures/events interrupted testing, thereby invali-dating the scores for that particular test. In other cases, some indi-viduals may have been unable to complete a given test andtherefore a valid score was not obtained. Missing data were im-puted using the EM maximum likelihood estimation procedure inthe EQS Structural Equations Program, which aims to impute miss-ing data while preserving the covariance matrix and sample means[36].

Multivariate multiple regression analyses were performed tomodel the relationship between the dependent measures (scoreson neuropsychological tests) and the following predictors: dura-tion of temporal lobe epilepsy/PNES (Duration), age at onset (Age-atOnset), education (Education), diagnostic group (Group), andlifetime seizure count (LifeFreq). Age-corrected raw scores of theexecutive functioning measures were used in the analyses,whereas scaled scores were used for the WAIS-III or WAIS-R devi-ation quotients.

In the initial multivariate multiple regression model, there weresignificant main effects for AgeatOnset, Education, and Duration.There was also a significant main effect for Group, suggesting arelationship between type of diagnosis (LTLE, RTLE, BiTLE, or PNES)

Table 1Demographic and seizure characteristics by diagnostic group.

Diagnostic group

LTLE RTLE BiTLE PNES

N 102 87 18 217Age (years) 37.07 37.07 40.89 36.67

(11.5) (10.63) (10.81) (10.79)Sex (% male) 51% 44% 33% 22%Education (years) 12.63 13.32 12.94 12.54

(2.52) (2.19) (2.75) (2.23)Full Scale IQ 91.76 92.74 88.61 91.56

(13.21) (12.34) (10.71) (11.82)AgeatOnset 20.12 18.38 19.69 28.99

(14.76) (13.58) (16.12) (12.81)Duration (years) 16.95 18.69 21.19 7.67

(13.34) (12.43) (15.77) (9.91)LifeFreq (No. of seizures) 3,238 3,953 3,368 5,594

(9685) (12,689) (4687) (33,972)

Note. Values are means (SD). LTLE, left temporal lobe epilepsy; RTLE, right temporallobe epilepsy; BiTLE, bilateral temporal lobe epilepsy; PNES, psychogenic nonepi-leptic seizures.

Table 2Descriptive statistics for transformed dependent variables.

Mean SD

WordFlua 34.23 11.09StroopCWa 37.74 10.80TrailsAa 3.42 0.40TrailsBa 4.28 0.42WCSTCata 4.71 1.80WCSTPera 2.58 0.74WMDQb 94.25 11.24VCDQ b 92.63 14.11PCDQ b 94.75 12.46

a Age-corrected raw score used.b Scaled score used.

414 L.C. Black et al. / Epilepsy & Behavior 17 (2010) 412–419

Black et al. (2010). The effect of seizures on working memory and executive function performance. Epilepsy & Behavior, 17, 412-419

ELT n=207 Crisis psicogénicas no epilépticas n=216

Análisis de regresión múltiple VI= edad de inicio, duración (años) y lateralidad (D, I, B) buenos predictores del desempeño neuropsicológico

WAIS-III

and performance. Education accounted for 33% of the variance inthe model. As education plays an integral part in performance onmany neuropsychological tests, especially IQ tests [37], it was ex-cluded from further analysis to allow focus on more clinically rel-evant predictors. The omnibus test of whether all of the R2 valueswere zero was significant, Wilk’s K = 0.53, F(45, 1837) = 6.14,P < 0.0001, indicating a significant relationship between the groupof dependent variables and the set of predictor variables. Theresulting model (summarized in Table 3), which excluded Educa-tion, showed a main effect of AgeatOnset, Group, and Duration.The omnibus test of whether all of the R2 values were zero withoutEducation in the model was significant: Wilk’s K = 0.76,F(36, 1542) = 3.18, P < 0.0001. As AgeatOnset and Duration in-creased, scores on many of the neuropsychological tests worsened.

Additional analyses, not reported here, showed no significantmultivariate differences among the three TLE groups. Therefore, toevaluate the effect of TLE versus PNES, further analyses were donecollapsing LTLE, RTLE, and BiTLE into one TLE group. Moderated uni-variate regression analyses were conducted in which three addi-tional predictors were added to each model shown in Table 3, thethree being centered interaction terms created by multiplying thedummy variable for Group (coded 0 for TLE and 1 for PNES) withthe deviation of each of the three continuous predictors from theirrespective means. If the incremental change in R2 that resulted fromthese three added predictors was significant, it was evidence that theregressive relationship varied as a function of group. Six of thenine univariate models showed significant evidence of moderation:WordFlu (R2 = 0.032), F(6, 417) = 2.27, P = 0.036; StroopCW (R2 =0.037), F(6, 417) = 2.64, P = 0.016; WCSTCat (R2 = 0.035),F(6, 417) = 2.53, P = 0.021; VCDQ (R2 = 0.082), F(6, 417) = 6.20,P < 0.001; TrailsA (R2 = 0.049), F(6, 417) = 3.55, P = 0.002; and TrailsB

(R2 = 0.077), F(6, 417) = 5.78, P < 0.001. PCDQ (R2 = 0.029) withF(6, 417) = 2.09, P = 0.053, approached significance.

To understand the nature of the moderated models, we con-ducted separate regressions for the PNES and TLE groups. Tables4 and 5 summarize the models for the PNES and TLE groups,respectively. Within the PNES sample, the independent variablesof AgeatOnset, Duration, and LifeFreq accounted for significant var-iation in one or more of the following dependent variables: WCST-Cat, PCDQ, TrailsA, and TrailsB (see Table 4 for univariate values forthe PNES group). Within the TLE sample, the independent variablesof AgeatOnset and LifeFreq accounted for significant variation inone or more of the following dependent variables: StroopCW,VCDQ, TrailsA, and TrailsB (see Table 5 for univariate values forthe TLE group). On the StroopCW, increased number of lifetime sei-zures resulted in fewer word colors correctly identified in the TLEgroup (b = –2.98) than in the PNES group (b = –1.70). On the WCST-Cat, as age at disorder onset became earlier in the PNES group, thenumber of categories achieved improved (b = –0.04), whereas theopposite was true in the TLE group (b = 0.02). On the VCDQ, laterage at disorder onset resulted in better performance in the TLEgroup (b = 0.41) than in the PNES group (b = 0.15). On the PCDQ,the longer the individual had the disorder, the better his or her per-formance was in the PNES group (b = 0.96), but the worse his or herperformance was in the TLE group (b = –0.25). On TrailsA, in-creased number of lifetime seizures resulted in longer latenciesin both the PNES (b = 0.09) and TLE (b = 0.11) groups. On TrailsB,increased number of lifetime seizures resulted in longer latenciesin both the PNES (b = 0.13) and TLE (b = 0.10) groups. Also onTrailsB, the earlier the disorder onset, the longer the latency tocomplete the trail in the PNES group (b = 0.01) but to a lesser ex-tent in the TLE group (b = 0.003).

Table 3Multivariate and univariate regression model results for dependent and predictor variables.

Multivariate results

Total model Group AgeatOnset (years) Duration LifeFreq (No. of seizures)

Wilks’ K 0.76 0.90 0.89 0.92 0.97F 3.18 1.68 5.67 3.82 1.29df 36,1542 27,1195 9,409 9,409 9,409P <0.001a 0.02a 0.00a 0.00a 0.24Power 0.99 1.00 0.99 0.63

b Weight

Intercept Group (relative to PNES) AgeatOnset (years) Duration LifeFreq (No. of seizures)

LTLE RTLE BiTLE

WordFlu 36.39 –3.31a –0.17 –0.74 0.01 –0.64 –0.05StroopCW 40.40 3.36a 4.92a –1.09 –0.01 –2.09a 0.03WCSTCat 5.41 0.28 0.09 0.23 –0.01 –0.21 –0.01WCSTPer 2.54 –0.06 –0.01 0.05 0.0001 0.06 –0.02WMDQ 92.15 1.62 2.77 –0.74 0.10 0.42 –0.44VCDQ 83.14 –1.62 2.31 –1.48 0.30a 1.90 –0.38PCDQ 89.25 3.30a 2.33 –2.25 0.11 –0.19 0.45TrailsA 3.11 –0.15 –0.07 0.19 0.004a 0.10a 0.01TrailsB 3.87 –0.01 –0.02 0.16 0.01a 0.12a 0.005

Univariate results

R2 F df P

WordFlu 0.03 1.93 6417 0.08StroopCW 0.05 3.54 6417 0.002WCSTCat 0.01 0.83 6417 0.55WCSTPer 0.01 0.43 6417 0.86WMDQ 0.02 1.72 6417 0.12VCDQ 0.07 5.21 6417 <0.0001a

PCDQ 0.02 1.71 6417 0.12TrailsA 0.06 4.4 6417 <0.0001a

TrailsB 0.06 4.74 6417 <0.0001a

a P < 0.05.

L.C. Black et al. / Epilepsy & Behavior 17 (2010) 412–419 415

32 D.D. Wagner et al. / Neuropsychologia 47 (2009) 112–122 117

or of side of epilepsy [F(1,83) = 0.62, p = 0.43]. There was a maineffect of group [F(3,83) = 3.1, p = 0.03, Munoperated = 0.39; MSAH = 0.57;MCAH = 0.51; Mtumorectomy = 0.23]: patients with SAH made moreintrusion errors overall than unoperated patients [t(49) = 2.53,p = 0.013]. The interaction between material type and side ofepilepsy was significant [F(1,83) = 7.74, p = 0.008]. Patients withleft TLE made more intrusion errors on the verbal supraspantask than patients with right TLE [unequal variances assumedt(88) = 2.81, p = 0.006 Mleft = 0.58; Mright = 0.40]. There was no simi-lar difference in intrusion errors on the visuospatial supraspan task[t(90) = 1.23, p = 0.221, Mleft = 0.45; Mright = 0.53]. The interactions ofpatient group and side of epilepsy; material type and patient group;and material type, patient group and side of epilepsy were all n.s.(p > 0.23, all tests).

3.2.2. Supraspan intrusion errors in left and right TLE and HCMixed-design ANOVA revealed no main effect of material type

[F(1,117) = 0.04, p = 0.847] or of side [F(2,117) = 2.3, p = 0.105]. Therewas a significant interaction between material type and group[F(2,117) = 5.09, p = 0.008] (Fig. 3). As in the previous analysis, thisinteraction indicated that patients with left TLE made more intru-sion errors than did right TLE patients on the verbal supraspan task[unequal variances assumed t(88) = 2.81, p = 0.006] but not on thevisuospatial task [t(90) = 1.23, p = 0.221]. Left TLE patients also mademore intrusion errors than healthy control subjects on the verbalsupraspan task [t(82) = 2.46, p = 0.016] but not on the visuospatialtask [t(82) = 0.87, p = 0.387]. Right TLE patients, on the other hand,made more intrusion errors than did control subjects on the visu-ospatial supraspan task [t(66) = 2.06, p = 0.043] but not on the verbaltask [t(64) = 0.16, p = 0.876]. Finally, left TLE patients made moreintrusion errors on the verbal supraspan task than on the visu-ospatial supraspan task [t(53) = 2.56, p = 0.013], but right TLE andHC showed no differences between material types on this measure[right TLE: t(35) = 1.96, p = 0.058; HC: t(29) = 0.18, p = 0.859].

3.3. Predictors of supraspan size and intrusion errors

Results of the multiple regression analysis examining demo-graphic and clinical variables that predicted verbal supraspan sizein patients with left or right TLE resulted in a single-predictormodel accounting for 12.5% of the variance in verbal supraspan

Fig. 2. Performance on verbal and visuospatial supraspan tasks as a function ofgroup. Left and right TLE groups are comprised of both unoperated and operatedpatients. Error bars indicate standard error of the mean.

Fig. 3. Ratio of intrusion errors to supraspan set size on verbal and visuospatialsupraspan tasks as a function of group. Left and right TLE groups are comprised ofboth unoperated and operated patients. Error bars indicate standard error of themean.

size [F(6,89) = 3.27, p = 0.006, R2 = 0.181, R2adj = 0.125]. Age was a sig-

nificant predictor of verbal supraspan size [ˇ = −0.25, p = 0.022],demonstrating that increasing age led to a decrease in verbalsupraspan size. Multiple regression analysis of factors predict-ing visuospatial supraspan size yielded a three-predictor modelaccounting for 18% of the variance in visuospatial supraspan size[F(6,89) = 4.48, p = 0.001, R2 = 0.232, R2

adj = 0.18]. The three variablesthat were significant predictors of visuospatial supraspan size wereage [ˇ = −0.342, p = 0.001], Full-Scale IQ [ˇ = 0.251, p = 0.023] andside of TLE [ˇ = −0.218, p = 0.028], indicating that increasing agepredicts a decline in visuospatial supraspan size and that rightTLE predicts a reduced visuospatial supraspan size compared toleft TLE—replicating the results found in the above mixed-designANOVAs. Finally, increasing Full-Scale IQ was predictive of a largervisuospatial supraspan size. The only common predictor acrossmaterial type was age.

Identical multiple regression analyses were conducted on intru-sion error ratios for both tasks. The model for verbal intrusion errorratios was significant, accounting for 11% of the variance in intru-sion errors [F(6,89) = 2.86, p = 0.013, R2 = 0.162, R2

adj = 0.105] andyielding a three-factor model with surgery [ˇ = 0.229, p = 0.025],side of epilepsy [ˇ = −0.227, p = 0.029] and Full-Scale IQ [ˇ = −0.231,p = 0.045] as significant predictors. This in effect replicates the pre-vious mixed-design ANOVA, revealing that operated patients mademore intrusion errors than unoperated patients and patients withleft TLE made more intrusion errors than those with right TLE onthe verbal task. Furthermore, Full-Scale IQ emerged as a significantpredictor of intrusion errors, indicating that patients with higherFull-Scale IQs were less likely to make intrusion errors on the ver-bal task. The model for visuospatial intrusion error ratios was notsignificant [F(6,89) = 1.83, p = 0.102], R2 = 0.110, R2

adj = 0.05].

3.4. Supraspan size with age as covariate

Results from the regression model consistently show that agenegatively predicts both verbal and visuospatial supraspan sizein patients. Furthermore, as pointed out above, right TLE patientswere older than left TLE patients [t(94) = 2.5, p = 0.014]. Giventhat our right TLE group is, on average, older than our left TLEgroup, it is possible that the effects seen in the mixed-design

Dígitos Cubos de Corsi

Controles sanos

Wagner et al. (2009). Material specific lateralization of WM in the medial temporal lobe epilepsy. Neuropsychologia, 47, 112-122

33

Remote Effects of Hippocampal Sclerosis on Effective Connectivity during WorkingMemory Encoding: A Case of Connectional Diaschisis?

Pablo Campo1,2, Marta I. Garrido3, Rosalyn J. Moran3, Fernando Maestu2, Irene Garcıa-Morales4,5, Antonio Gil-Nagel5,Francisco del Pozo2, Raymond J. Dolan3 and Karl J. Friston3

1Department of Basic Psychology, Autonoma University of Madrid, 28049 Madrid, Spain, 2Center for Biomedical Technology,Laboratory of Cognitive and Computational Neuroscience, Complutense University of Madrid–Polytechnic University of Madrid,28223 Madrid, Spain, 3Institute of Neurology, Wellcome Trust Centre for Neuroimaging, University College London, LondonWC1N3BG, UK, 4Epilepsy Unit, Department of Neurology, University Hospital of San Carlos, 28040 Madrid, Spain and 5Epilepsy Unit,Department of Neurology, Hospital Ruber Internacional, 28034 Madrid, Spain

Address correspondence to Dr Pablo Campo, Departamento de Psicologıa Basica, Universidad Autonoma de Madrid, Campus de Cantoblanco, 28049Madrid, Spain. Email: [email protected].

Accumulating evidence suggests a role for the medial temporal lobe(MTL) in working memory (WM). However, little is knownconcerning its functional interactions with other cortical regionsin the distributed neural network subserving WM. To reveal these,we availed of subjects with MTL damage and characterizedchanges in effective connectivity while subjects engaged in WMtask. Specifically, we compared dynamic causal models, extractedfrom magnetoencephalographic recordings during verbal WMencoding, in temporal lobe epilepsy patients (with left hippocampalsclerosis) and controls. Bayesian model comparison indicated thatthe best model (across subjects) evidenced bilateral, forward, andbackward connections, coupling inferior temporal cortex (ITC),inferior frontal cortex (IFC), and MTL. MTL damage weakenedbackward connections from left MTL to left ITC, a decreaseaccompanied by strengthening of (bidirectional) connectionsbetween IFC and MTL in the contralesional hemisphere. Thesefindings provide novel evidence concerning functional interactionsbetween nodes of this fundamental cognitive network and shedslight on how these interactions are modified as a result of focaldamage to MTL. The findings highlight that a reduced (top-down)influence of the MTL on ipsilateral language regions is accompa-nied by enhanced reciprocal coupling in the undamaged hemisphereproviding a first demonstration of ‘‘connectional diaschisis.’’

Keywords: dynamic causal modeling, effective connectivity,magnetoencephalography, temporal lobe epilepsy, working memory

Introduction

Extensive evidence indicates that medial temporal lobe (MTL) isnot exclusively involved in long-term memory (LTM). Humanneuroimaging studies have reported activation of MTL duringworking memory (WM) tasks that engage informational encod-ing (Campo et al. 2005; Karlsgodt et al. 2005; Mainy et al. 2007),maintenance of information (Ranganath and D’Esposito 2001;Axmacher et al. 2007), and retrieval (Cabeza et al. 2002; Schonet al. 2009). Also supporting this view, neuropsychological andneuroimaging studies have revealed impaired performance andabnormalities in MTL activity during WM tasks in patients withMTL damage with various causes (Owen et al. 1996; Krauss et al.1997; Abrahams et al. 1999; Grady et al. 2001; Lancelot et al.2003; Lee et al. 2006; Olson et al. 2006; Piekema et al. 2007;Ezzyat and Olson 2008; Wagner et al. 2009). However, based onthe assumption that cognitive processes engage distributedneural networks, if we want to gain a clearer understanding of

the functional role of MTL in WM it cannot be considered as anindependent processor. It is, therefore, necessary to character-ize that role from the perspective of the functional systems(Bullmore and Sporns 2009). Accordingly, the goal of thecurrent study was to investigate the conjoint function of MTLand other functionally related brain regions involved in verbalWM as a large-scale network (Bressler and Menon 2010). Weobtained whole-head magnetoencephalographic (MEG) record-ings during a verbal WM task, which was designed to ensure thatparticipants encoded words semantically (Campo et al. 2005,2009), as prior neuroimaging investigations have demonstratedthat depth processingmodulates MTL activity (Kapur et al. 1994;Lepage et al. 2000).

Although few previous studies have used connectivityanalyses to investigate the interactions between MTL and otherkey structures in the neural network involved in visual andverbalWM (Petersson et al. 2006; Nee and Jonides 2008; Rissmanet al. 2008), our study diverges from those in 2 main aspects.First, we studied temporal lobe epilepsy (TLE) patients with lefthippocampal sclerosis (HS) (Trenerry et al. 1993; Thom et al.2005) in order to evaluate the impact of unilateral MTLpathology on functional organization and connectivity amongbrain regions engaged in verbal WM encoding. This isconsidered as a useful approach that allows the characterizationof changes in the functional organization of interconnectedbrain regions following focal brain damage (Guye et al. 2008).Second, we used dynamic causal modeling (DCM) (Friston et al.2003; David et al. 2006; Daunizeau et al. forthcoming) tocharacterize the effective connectivity in the WM network, insubjects with and without MTL damage (Seghier et al. 2010).Effective connectivity denotes ‘‘directed or causal relationshipsbetween elements’’ (Bullmore and Sporns 2009) and in thepresent context refers to the change that the activity in onebrain region causes in the activity of another, and how this ismodulated by experimental factors (Stephan and Friston 2007).Effective connectivity can be estimated with Bayesian modelinversion by perturbing the system and measuring its response(Friston and Price 2001; Garrido, Kilner, Kiebel, and Friston2007)–this is DCM. DCM represents a fundamental variationfrom alternative methods to estimate connectivity because itemploys a generative model of measured brain responses thattakes into account their nonlinear and dynamic nature. Asopposed to functional connectivity measures that explorenondirectional statistical dependencies between brain regions,

! The Authors 2011. Published by Oxford University Press.This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/2.5), which permitsunrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Cerebral Cortex June 2012;22:1225–1236Cerebral Cortex June 2012;22:1225–1236doi:10.1093/cercor/bhr201doi:10.1093/cercor/bhr201Advance Access publication August 1, 2011

by guest on October 2, 2012

http://cercor.oxfordjournals.org/D

ownloaded from







Introduction








ownloaded from

34

Effective Connectivity Analysis: DCMDetermining effective connectivity requires a causal model of theinteractions among the constituents of the neural network subject tostudy (Stephan and Friston 2007). DCM considers the brain as ‘‘adeterministic nonlinear dynamical system that is subject to inputs andproduces outputs’’ (David 2007).DCM is a hypothesis-driven method that relies on the specification of

a plausible biophysical and physiological model of interacting brainregions (Stephan and Friston 2007). The model is specified by itsregions connections and by whether these connections are unidirec-tional (forward or backward) or bidirectional (both forward andbackward). Forward and backward connections are defined accordingto the connectivity rules outlined in Felleman and Van Essen (1991)and specified in DCM to convey bottom-up and top-down effects,respectively. This model is then supplemented with a forward model ofhow neuronal or synaptic activity is transformed into a measuredresponse (Kiebel et al. 2006). This enables the parameters of theneuronal model (i.e., effective connectivity) and spatial model (i.e.,dipole orientations) to be estimated from observed data using a Bayesianscheme. Estimating the parameters of a DCM model relies on estimatingthe hidden states and parameters of the modeled system, whichcorresponds to the sources that comprise the model (David et al.2006). DCM for MEG uses a neural mass model to explain sourceactivity (David and Friston 2003) and has been described in detailelsewhere (David et al. 2006).

DCM Specification: Hypotheses TestedNetwork architecture was specified on the basis of the inversesolutions (source localizations; see Fig. 1) for single subjects usingmultiple sparse priors (Friston et al. 2008) and was constrained byrecent studies of functional connectivity on verbal WM (Fiebach et al.2006; Nee and Jonides 2008). Accordingly, we considered for ourmodels 6 regions that corresponded to ITC, MTL, and VLPFC/IFCbilaterally. These sources were modeled as equivalent current dipoles,which were superimposed on an MRI of a standard brain in MNI space(Fig. 1), whose prior mean locations coordinates (x, y, z) are: bilateralITC: –43, –54, –15 (left); 43, –54, –15 (right); bilateral MTL: –27, –15, –20(left); 27, –15, –20 (right); and bilateral IFC/VLPFC: –54, 35, 6 (left); 54,

35, 6 (right). Twelve models were specified and inverted separately foreach subject (Fig. 2b). In all models, left and right ITC were chosen asvisual input nodes for semantic processing of words (Bitan et al. 2005;Heim et al. 2009). The models were specified starting with simplearchitectures and adding hierarchical levels (i.e., sources and extrinsicconnections). The simplest models only included the ITC and IFC/VLPFC sources, while more complex models included MTL sources.The sources were left unilateral, right unilateral, or bilateral. Modelsalso differed in terms of their connections; forward only or bothforward and backward. Accordingly, model lF – included left unilateraland forward connections, while model bFB – included bilateral sourceswith forward and backward connections. Models with MTL sourceswere created by simply adding MTL sources; that is, model lF+ includedleft unilateral and forward connections and the MTL. See Figure 2b, fordetails. The MTL models allowed an evaluation of the involvement ofMTL in verbal WM and the functional relevance of the connections ofthis region within the network.

Model ComparisonOne of the advantages of DCM is that it can be used to comparecompeting hypotheses about functional architectures (David 2007;Garrido, Kilner, Kiebel, Stephan, et al. 2007; Friston 2009; Garrido,Kilner, Kiebel, Stephan, et al. 2009). This is accomplished by specifyinga model (hypothesis), in terms of anatomical connections betweenbrain regions. Using Bayesian model selection, DCM tests a group ofcompeting models and provides evidence in favor of one model,relative to others (Penny et al. 2004). The model log-evidence orthe marginal log-likelihood of each model is compared against theremaining models. The model with the highest evidence (i.e., the modelwith the best balance of accuracy and complexity) is then consideredthe best or optimal model. A difference of 3 or more in favor of onemodel as compared with others is required (Penny et al. 2004). Weperformed a fixed-effect analysis for comparing model log-evidence atthe group level (i.e., patient group and control group), which isaccomplished by summing the log-evidence of each participant foreach model, finding the highest valued model and comparing it withthe summed log evidence of the next highest model (Garrido, Kilner,Kiebel, and Friston 2007; Garrido, Kilner, Kiebel, Stephan, et al. 2009;

Figure 1. Source localization for a representative subject using multiple sparse priors (upper panel). Sources of activity, modeled as dipoles (estimated posterior moments andlocations) superimposed in an MRI of a standard brain in MNI space (lower panel).

Cerebral Cortex June 2012, V 22 N 6 1227



ownloaded from







Introduction








ownloaded from

35







Introduction








ownloaded from

Figure 3. (A) Group level Bayesian selection of the 12 tested models. Left: fixed effect analysis (FFX) showing log-evidence and model posterior probability. Right: random fixedeffects (RFX) showing model expected probability and model exceedance probability. Results indicate the best model is one with bilateral forward and backward connectionscomprising IFG, ITC, and MTL. (Bayes factor relative to the second best model [model lFBþ] 5 452.07; exceedance probability for model bFBþ 5 0.965). 1. LF "; 2. RF "; 3. BF"; 4. LFB "; 5. RFB "; 6. BFB "; 7. LF þ; 8. RF þ; 9. BF þ; 10. LFB þ; 11. RFB þ; 12. BFB þ. L, left; R, right; B, bilateral; F, forward; FB, forward and backward; " modelarchitecture not including MTL; þ model architecture including MTL. (B) Predicted (blue) and observed (red) responses in measurement space for the best model. (C) Groupdifferences in effective connectivity assessed using subject-specific (maximum a posteriori) parameter estimates.

Connectivity Changes in WM Network d Campo et al.1230



ownloaded from

Conexiones LTM y LTI

Conexiones LTM y LFI contralateral

Déficit de memoria episódica (autobiográfica)

• Memoria normal o casi a los 30 min., pero alterada en periodos más largos (días o semanas) i.e. alteración de los mecanismos de consolidación.

•  Las alteraciones de memoria se han estudiado en la Amnesia Epiléptica Transitoria (TEA).

• Butler et al. (2007) 50 pacientes con TEA 70% reportaron déficits de la memoria autobiográfica que podía extenderse varias décadas.

36

BRAINA JOURNAL OF NEUROLOGY

Remote memory deficits in transient epilepticamnesiaFraser Milton,1 Nils Muhlert,2 Dominika M. Pindus,1 Christopher R. Butler,3 Narinder Kapur,4

Kim S. Graham5 and Adam Z. J. Zeman1,2

1 School of Psychology, University of Exeter, Exeter, EX4 4QG, UK2 Peninsula Medical School, University of Exeter, EX1 2LU, UK3 Department of Clinical Neurology, University of Oxford, Oxford, OX3 9DU, UK4 Neuropsychology Department, Addenbrooke’s Hospital, Cambridge, CB2 0QQ, UK5 Wales Institute of Cognitive Neuroscience, School of Psychology, Cardiff University, Cardiff, CF 10 3AT, UK

Correspondence to: Fraser Milton,Washington Singer Laboratories,Perry Road, Exeter,UK, EX4 4QGE-mail: [email protected]

Transient epileptic amnesia is a form of temporal lobe epilepsy in which sufferers often complain of irretrievable loss of remote

memories. We used a broad range of memory tests to clarify the extent and nature of the remote memory deficits in patients

with transient epileptic amnesia. Performance on standard tests of anterograde memory was normal. In contrast, there was a

severe impairment of memory for autobiographical events extending across the entire lifespan, providing evidence for the

occurrence of ‘focal retrograde amnesia’ in transient epileptic amnesia. There was a milder impairment of personal semantic

memory, most pronounced for midlife years. There were limited deficits of public semantic memory for recent decades. These

results may reflect subtle structural pathology in the medial temporal lobes or the effects of the propagation of epileptiform

activity through the network of brain regions responsible for long-term memory, or a combination of these two mechanisms.

Keywords: transient epileptic amnesia; remote memory; autobiographical memory; focal retrograde amnesia

Abbreviations: TEA = transient epileptic amnesia

IntroductionMemory complaints are common among people with epilepsy

(Corcoran and Thompson, 1992), especially among patients with

temporal lobe epilepsy in which key structures involved in process-

ing memories, including the hippocampus, are directly involved by

seizure activity (Butler and Zeman, 2008a). However, whilst there

is extensive evidence for anterograde memory deficits in temporal

lobe epilepsy, relatively few studies have investigated remote

memory (Noulhiane et al., 2007; Butler and Zeman, 2008a).

Nevertheless, remote memory deficits can have considerable

impact on psychological well-being and are sometimes the

presenting feature of patients with temporal lobe epilepsy

(Gallassi, 2006).

Remote memory is multi-faceted, comprising memories that

were encoded in the relatively distant past, arbitrarily defined as

over one year ago (Kapur, 1999; Butler and Zeman, 2008a).

Remote memory has episodic and semantic components.

Episodic memory is typically autobiographical, involving the recol-

lection of personally experienced events and allowing ‘mental time

travel’ into the past, or ‘autonoetic awareness’ (Tulving, 1985).

Semantic memory enables the recollection of declarative facts

doi:10.1093/brain/awq055 Brain 2010: 133; 1368–1379 | 1368

Received November 6, 2009. Revised February 14, 2010. Accepted February 15, 2010. Advance Access publication March 31, 2010! The Author (2010). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved.For Permissions, please email: [email protected]

by guest on March 26, 2015

Dow

nloaded from

scoring higher than patients. T-tests indicated that controls scored

significantly higher than patients for all Time periods (P50.01).

A post-memory retrieval rating indicated that there was no dif-

ference between patients (mean = 3.14, SD = 1.19) and controls

(mean = 3.46, SD = 0.72) in the personal significance of the

memories [t(24) = 0.81, P = 0.43].

Crovitz TestThe memories provided by controls (mean = 26.50, SD = 2.07)

scored significantly higher than those produced by patients

(mean = 20.64, SD = 4.67) [t(24) = 4.01, P = 0.001]. An

ANOVA (Time ! Group) assessed differences between patients

and controls in the distribution of memories over time (Fig. 3).

This yielded a significant effect of time [F(3,72) = 12.82,

P50.001] indicating a bias toward retrieving more recent mem-

ories. There was a significant interaction between Time period and

Group [F(3,72) = 24.69, P = 0.031]. Pairwise comparisons indi-

cated that patients retrieved significantly fewer memories than

controls from the youth period (P50.05) but produced more

from the most recent period, although this effect missed signifi-

cance (P = 0.053).

Personal semantic memoryFigure 4 shows personal semantic memory performance across

time for both groups. An ANOVA (Time period ! Group) revealed

an effect of Group [F(1,24) = 9.50, P = 0.005] indicating that pa-

tients recalled significantly fewer personal semantic details than

controls. There was a marginally significant effect of Time

[F(4,96) = 2.47, P = 0.05] but no interaction between Time and

Group [F(4,96) = 1.89, P = 0.12]. T-tests revealed that controls

recalled significantly more personal semantic details than patients

for the middle-age period (P = 0.002); the remaining periods were

not significant (P40.1).

Public semantic tests

Dead-or-Alive Test

Table 4 shows the mean performance for patients and controls in

the Dead-or-Alive Test measures. T-tests revealed that controls

0

5

10

15

20

25

Time period

Inte

rnal

det

ails

at r

ecal

l

0

5

10

15

20

25

30

35

40

Time period

Ext

erna

l det

ails

at r

ecal

l

0

5

10

15

20

Child Youth Youngadult

Middle age Recent


Middle age Recent


Middle age Recent

Time period

Rec

all r

atin

gs

******

**

Patients Controls

A

B

C

Figure 1 (A) Mean number of internal details recalled for eachtime period at recall; (B) mean number of external details re-called for each time period at recall; (C) mean rating (out of 21)for each time period at recall. *P50.05; *** P50.005.

Table 3 Performance on anterograde memory tests for patients with TEA and controls

Anterograde memory scores (max score) TEA groupMean (SD)

Control groupMean (SD)

P-value

Episodic memory scores

Story recall immediate (25) 14.00 (1.88) 13.67 (5.25) 0.83

Story recall delayed (25) 12.21 (2.33) 11.75 (4.97) 0.76

Story recognition (15) 12.86 (1.70) 12.33 (2.10) 0.49

Visuospatial perception scores

Rey Complex Figure Copy (36) 33.14 (2.32) 32.96 (3.56) 0.88

Rey Complex Figure Delayed Recall (36) 16.89 (5.78) 16.71 (6.98) 0.94

Warrington Recognition Memory Test

Word recognition (50) 48.00 (1.83) 46.73 (3.80) 0.54

Face recognition (50) 42.00 (3.81) 43.00 (3.98) 0.30

Paired Associates Learning (units) "0.77 (1.85) 0.09 (0.80) 0.17

TEA and remote memory Brain 2010: 133; 1368–1379 | 1373


Dow

nloaded from

scoring higher than patients. T-tests indicated that controls scored

significantly higher than patients for all Time periods (P50.01).

A post-memory retrieval rating indicated that there was no dif-

ference between patients (mean = 3.14, SD = 1.19) and controls

(mean = 3.46, SD = 0.72) in the personal significance of the

memories [t(24) = 0.81, P = 0.43].

Crovitz TestThe memories provided by controls (mean = 26.50, SD = 2.07)

scored significantly higher than those produced by patients

(mean = 20.64, SD = 4.67) [t(24) = 4.01, P = 0.001]. An

ANOVA (Time ! Group) assessed differences between patients

and controls in the distribution of memories over time (Fig. 3).

This yielded a significant effect of time [F(3,72) = 12.82,

P50.001] indicating a bias toward retrieving more recent mem-

ories. There was a significant interaction between Time period and

Group [F(3,72) = 24.69, P = 0.031]. Pairwise comparisons indi-

cated that patients retrieved significantly fewer memories than

controls from the youth period (P50.05) but produced more

from the most recent period, although this effect missed signifi-

cance (P = 0.053).

Personal semantic memoryFigure 4 shows personal semantic memory performance across

time for both groups. An ANOVA (Time period ! Group) revealed

an effect of Group [F(1,24) = 9.50, P = 0.005] indicating that pa-

tients recalled significantly fewer personal semantic details than

controls. There was a marginally significant effect of Time

[F(4,96) = 2.47, P = 0.05] but no interaction between Time and

Group [F(4,96) = 1.89, P = 0.12]. T-tests revealed that controls

recalled significantly more personal semantic details than patients

for the middle-age period (P = 0.002); the remaining periods were

not significant (P40.1).

Public semantic tests

Dead-or-Alive Test

Table 4 shows the mean performance for patients and controls in

the Dead-or-Alive Test measures. T-tests revealed that controls

0

5

10

15

20

25

Time period

Inte

rnal

det

ails

at r

ecal

l

0

5

10

15

20

25

30

35

40

Time period

Ext

erna

l det

ails

at r

ecal

l

0

5

10

15

20


Middle age Recent


Middle age Recent


Middle age Recent

Time period

Rec

all r

atin

gs

******

**

Patients Controls

A

B

C

Figure 1 (A) Mean number of internal details recalled for eachtime period at recall; (B) mean number of external details re-called for each time period at recall; (C) mean rating (out of 21)for each time period at recall. *P50.05; *** P50.005.

Table 3 Performance on anterograde memory tests for patients with TEA and controls

Anterograde memory scores (max score) TEA groupMean (SD)

Control groupMean (SD)

P-value

Episodic memory scores

Story recall immediate (25) 14.00 (1.88) 13.67 (5.25) 0.83

Story recall delayed (25) 12.21 (2.33) 11.75 (4.97) 0.76

Story recognition (15) 12.86 (1.70) 12.33 (2.10) 0.49

Visuospatial perception scores

Rey Complex Figure Copy (36) 33.14 (2.32) 32.96 (3.56) 0.88

Rey Complex Figure Delayed Recall (36) 16.89 (5.78) 16.71 (6.98) 0.94

Warrington Recognition Memory Test

Word recognition (50) 48.00 (1.83) 46.73 (3.80) 0.54

Face recognition (50) 42.00 (3.81) 43.00 (3.98) 0.30

Paired Associates Learning (units) "0.77 (1.85) 0.09 (0.80) 0.17

TEA and remote memory Brain 2010: 133; 1368–1379 | 1373


Dow

nloaded from

Déficits en todos los contextos: •  Evento •  Tiempo •  Lugar •  Perceptivo •  Pensamiento/emoción

37

38

Review

Temporal lobe epilepsy as a model to understand human memory:The distinction between explicit and implicit memory

Elizabeth C. Leritz a,b,*, Laura J. Grande a,c, Russell M. Bauer d

a Geriatric Research, Education, and Clinical Center (GRECC), Boston VA Healthcare System, Boston, MA, USAb Division of Aging, Brigham & Women’s Hospital, Boston, MA, USA

c Department of Psychiatry, Harvard Medical School, Boston, MA, USAd Department of Clinical and Health Psychology, University of Florida, Gainesville, FL, USA

Received 4 October 2005; revised 10 March 2006; accepted 15 April 2006Available online 8 June 2006

Abstract

Decades of research have provided substantial evidence of memory impairments in patients with temporal lobe epilepsy (TLE),including deficits in the encoding, storage, and retrieval of new information. These findings are not surprising, given the associated under-lying neuroanatomy, including the hippocampus and surrounding medial temporal lobe structures. Because of its associated anatomicand cognitive characteristics, TLE has provided an excellent model by which to examine specific aspects of human memory functioning,including classic distinctions such as that between explicit and implicit memory. Various clinical and experimental research studies havesupported the idea that both conscious and unconscious processes support memory functioning, but the role of relevant brain structureshas been the subject of debate. This review is concerned with a discussion of the current status of this research and, importantly, howTLE can inform future studies of memory distinctions.! 2006 Elsevier Inc. All rights reserved.

Keywords: Temporal lobe epilepsy; Anterior temporal lobectomy; Explicit memory; Implicit memory; Associative memory; Relational memory;Hippocampus; Medial temporal lobe

1. Introduction

Memory impairment in patients with temporal lobe epi-lepsy (TLE) is well-documented. Due to the relatively cir-cumscribed nature of epilepsy-related pathology, whichtypically involves cell loss and gliosis in the hippocampusand surrounding structures in the medial temporal lobe,TLE has provided an excellent model by which to investi-gate specific aspects of the learning and retrieval of newinformation. This has allowed not only for a better under-standing of the role of medial temporal lobe structures inmemory, but has also provided a clearer picture of thememory deficits that specifically characterize patients withTLE. Of particular relevance has been the role that TLEhas played in helping to parse out the different systems that

have been posited to underlie human memory, and to fur-ther clarify the underlying anatomy. Establishing relation-ships between memory systems and underlying anatomyhas been a key feature of contemporary memory research.The current review is concerned with a classic distinction inthe memory literature, that between explicit and implicitmemory.

We begin with a brief review of common memory dis-tinctions, including a discussion of the basic themes under-lying each definition. Next, we review the literature onexplicit and implicit memory, including a discussion offindings in patient populations, and concluding with a sum-mary of the current status of the distinction and its rele-vance to memory research. We then briefly discuss brainstructures involved in TLE and their relation to the specificcognitive processes underlying both explicit and implicitmemory. Lastly, we specifically discuss the explicit andimplicit memory deficits that have been reported in TLE,

1525-5050/$ - see front matter ! 2006 Elsevier Inc. All rights reserved.doi:10.1016/j.yebeh.2006.04.012

* Corresponding author. Fax: +1 857 364 4544.E-mail address: [email protected] (E.C. Leritz).

www.elsevier.com/locate/yebeh

Epilepsy & Behavior 9 (2006) 1–13

Review

Temporal lobe epilepsy as a model to understand human memory:The distinction between explicit and implicit memory

Elizabeth C. Leritz a,b,*, Laura J. Grande a,c, Russell M. Bauer d

a Geriatric Research, Education, and Clinical Center (GRECC), Boston VA Healthcare System, Boston, MA, USAb Division of Aging, Brigham & Women’s Hospital, Boston, MA, USA

c Department of Psychiatry, Harvard Medical School, Boston, MA, USAd Department of Clinical and Health Psychology, University of Florida, Gainesville, FL, USA

Received 4 October 2005; revised 10 March 2006; accepted 15 April 2006Available online 8 June 2006

Abstract

Decades of research have provided substantial evidence of memory impairments in patients with temporal lobe epilepsy (TLE),including deficits in the encoding, storage, and retrieval of new information. These findings are not surprising, given the associated under-lying neuroanatomy, including the hippocampus and surrounding medial temporal lobe structures. Because of its associated anatomicand cognitive characteristics, TLE has provided an excellent model by which to examine specific aspects of human memory functioning,including classic distinctions such as that between explicit and implicit memory. Various clinical and experimental research studies havesupported the idea that both conscious and unconscious processes support memory functioning, but the role of relevant brain structureshas been the subject of debate. This review is concerned with a discussion of the current status of this research and, importantly, howTLE can inform future studies of memory distinctions.! 2006 Elsevier Inc. All rights reserved.

Keywords: Temporal lobe epilepsy; Anterior temporal lobectomy; Explicit memory; Implicit memory; Associative memory; Relational memory;Hippocampus; Medial temporal lobe

1. Introduction

Memory impairment in patients with temporal lobe epi-lepsy (TLE) is well-documented. Due to the relatively cir-cumscribed nature of epilepsy-related pathology, whichtypically involves cell loss and gliosis in the hippocampusand surrounding structures in the medial temporal lobe,TLE has provided an excellent model by which to investi-gate specific aspects of the learning and retrieval of newinformation. This has allowed not only for a better under-standing of the role of medial temporal lobe structures inmemory, but has also provided a clearer picture of thememory deficits that specifically characterize patients withTLE. Of particular relevance has been the role that TLEhas played in helping to parse out the different systems that

have been posited to underlie human memory, and to fur-ther clarify the underlying anatomy. Establishing relation-ships between memory systems and underlying anatomyhas been a key feature of contemporary memory research.The current review is concerned with a classic distinction inthe memory literature, that between explicit and implicitmemory.

We begin with a brief review of common memory dis-tinctions, including a discussion of the basic themes under-lying each definition. Next, we review the literature onexplicit and implicit memory, including a discussion offindings in patient populations, and concluding with a sum-mary of the current status of the distinction and its rele-vance to memory research. We then briefly discuss brainstructures involved in TLE and their relation to the specificcognitive processes underlying both explicit and implicitmemory. Lastly, we specifically discuss the explicit andimplicit memory deficits that have been reported in TLE,

1525-5050/$ - see front matter ! 2006 Elsevier Inc. All rights reserved.doi:10.1016/j.yebeh.2006.04.012

* Corresponding author. Fax: +1 857 364 4544.E-mail address: [email protected] (E.C. Leritz).

www.elsevier.com/locate/yebeh

Epilepsy & Behavior 9 (2006) 1–13

39

and we then provide some insights into future research onmemory distinctions in this patient population.

2. Memory distinctions

2.1. Brief overview

For years, memory researchers have relied on the use ofthe conceptual classification of memory distinctions to bet-ter understand normal and abnormal human memory func-tioning. Such distinctions have been particularly helpfulwhen characterizing particular neuropsychological disor-ders, in which impairment is often seen in one memory-re-lated domain but not in another. Consistent reports ofthese dissociations have led to the development of suchterms as declarative versus nondeclarative memory, declara-tive versus procedural memory, episodic versus semanticmemory, as well as explicit and implicit memory. Fig. 1 isa diagram containing brief descriptions of these classifica-tions and illustrating the theoretical relationships betweenthem.

The underlying conceptual themes behind each dissocia-tion overlap to some extent, and as such, many patientgroups who demonstrate deficits in one category also doso in another (designated in the figure by the depiction ofcategories with overlapping constructs). For example,under the rubric of declarative memory, episodic memoryand explicit memory both refer, in some degree, to theacquisition of new information, and memory-disorderedpatients typically have difficulty across tasks ascribed toeach of these domains. Implicit memory and proceduralmemory, as well as implicit memory and semantic memory,also contain similar definitions, and researchers often inter-change these terms when referring to learning under uncon-scious or ‘‘indirect’’ means (as in the case of implicitmemory and procedural memory), or when referring to

the access of memory stores that are not tied to a particularevent, place, or time (as in the case of semantic memory).In this review, we highlight the distinction between explicitand implicit memory, a dissociation that has received muchattention in neuropsychological research. Many experi-mental tasks have successfully dissociated performance indifferent patient populations, including neuropsychologicaldisorders such as amnesia and Alzheimer’s disease, as wellas developmental stages such as aging. Because of its directrelevance to patients with temporal lobe memory disorders,we focus solely on the class of implicit memory known asindirect memory, which includes tasks such as repetitionor associative priming (see Fig. 1). Often, researchers inter-change the term implicit memory with priming, which refersto the speeded processing of a stimulus as a result of priorexposure to or familiarity with that stimulus. We do notdiscuss procedural memory (also known as implicit skillor motor learning) or semantic memory as part of thereview.

2.2. Implicit and explicit memory: Definition of thedistinction

The distinction between explicit and implicit memoryhas been a dominant theme in memory research over thepast several decades. In this article, we define explicit mem-ory (also referred to as direct memory) as the intentionalrecollection of newly learned information. Thus, we referto the recall of episodic information, such as that whichis acquired during the study, or learning, phase of a mem-ory experiment. Critical to this widely used definition is theidea that explicit memory involves the retrieval of materialthat has been recently introduced to memory stores. Suchinformation may include facts and specific events, but doesnot include information that is not tied to a specific contextor that is considered ‘‘general knowledge’’ (semantic mem-

Fig. 1. Classification of commonly studied memory categories and distinctions. Overlapping circles indicate partially overlapping constructs underlyingeach category.

2 E.C. Leritz et al. / Epilepsy & Behavior 9 (2006) 1–13

Explicit

40

likely that functions underlying implicit tasks are more het-erogeneous. However, the robustness of the distinctionbetween explicit and implicit memory is fairly powerful,and has been observed across several patient groups,including amnesics, patients with early Alzheimer’s disease[65], aging patients [66], and patients with epilepsy [67], allwho commonly share damage to MTL structures as adefining feature.

Interestingly, there is no evidence of a reverse dissociationin patients with MTL damage; to our knowledge, there is noexisting report of impaired implicit but intact explicit mem-ory, in the context of the definitions we describe here. That is,there are no findings suggesting impaired repetition primingfor relational information in the face of intact explicit mem-ory for this same material. Thus, inferences of a distinctionbetween explicit and implicit memory, and between implicitmemory for preexisting material and that for novel material,are based solely on single dissociations, which are consider-ably less powerful than double dissociations [68]. For exam-ple, evidence of impaired priming of single-item stimuli in theface of intact memory for associative information wouldargue more strongly for the view that a true cognitive andneuroanatomic distinction exists between these mnemonicproperties. This has important implications when arguingfor and using the explicit/implicit distinction in both experi-mental and clinical research. In an effort to integrate thevarying results across studies and across patient populations,we propose a model that explains our view of how differentcomponents contribute to various aspects of explicit and

implicit memory. In addition, this model provides an illus-tration of why only single dissociations are observed whenevaluating this distinction (see Fig. 2).

We depict explicit and implicit memory on a continuumof dependence on the hippocampus, as well as of the degreeof effort/intentionality underlying each task. To make thisdepiction clearer, we also list tasks that lie at differentpoints on the continuum. It should be noted that withineach task, the level of complexity may vary with respectto stimuli that either are novel or require a more associativecomponent (as in an unrelated word pair), compared withstimuli that are item-specific or are theorized to have prees-tablished representations in memory; relational informa-tion is more difficult to encode and remember than singleitems. According to the overall model, hippocampalinvolvement is less critical as the continuum progressesfrom explicit to implicit memory. Thus, the hippocampalcomplex is most important for the explicit, intentionalrecall of complex information (top of the model, for exam-ple, free recall of previously learned unrelated word pairs),but becomes less taxed as retrieval is less effortful and more‘‘indirect’’ (bottom of the model, for example, tasks thatare based on memory for preexisting, item-specific infor-mation such as repetition priming of single words and mereperceptual priming). The model provides an illustration tosuggest that the hippocampal complex is important, albeitto varying degrees, for the laying down of new memories,regardless of the direct or indirect nature of the task. Thismodel is in contrast to views stemming from decades of

Fig. 2. Hypothesized continuum of explicit and implicit memory as it relates to hippocampal involvement and degree of effort and intentionality. In thismodel, hippocampus refers collectively to the hippocampal complex, which includes the hippocampus and surrounding MTL structures and cortices(parahippocampal gyrus, preirhinal and entorhinal cortex).

6 E.C. Leritz et al. / Epilepsy & Behavior 9 (2006) 1–13

•  Reducir los déficits cognitivos 1

•  Reducir los efectos adversos de los déficits 2 •  Mejorar la conciencia del déficit para que

puedan afrontar mejor los retos de la vida diaria

3 •  Incidir sobre otras areas no directamente

tratadas: depresión, ansiedad, fatiga, cambios de personalidad, calidad de vida

4

Propósito de la rehabilitación neuropsicológica

41

42 Journal of the International Neuropsychological Society (2014), 20, 868–872.Copyright © INS. Published by Cambridge University Press, 2014.doi:10.1017/S1355617714000630

BRIEF COMMUNICATION

Working Memory Mediates the Relationship between IntellectualEnrichment and Long-Term Memory in Multiple Sclerosis:An Exploratory Analysis of Cognitive Reserve

Joshua Sandry AND James F. SumowskiNeuropsychology and Neuroscience Laboratory, Kessler Foundation, West Orange, NJ, and Department of Physical Medicine and Rehabilitation,Rutgers – New Jersey Medical School

(RECEIVED February 21, 2014; FINAL REVISION June 17, 2014; ACCEPTED June 18, 2014; FIRST PUBLISHED ONLINE July 14, 2014)

Abstract

Some individuals with multiple sclerosis (MS) show decrements in long-term memory (LTM) while other individuals donot. The theory of cognitive reserve suggests that individuals with greater pre-morbid intellectual enrichment are protectedfrom disease-related cognitive decline. How intellectual enrichment affords this benefit remains poorly understood. Thepresent study tested an exploratory meditational hypothesis whereby working memory (WM) capacity may mediate therelationship between intellectual enrichment and verbal LTM decline in MS. Intellectual enrichment, verbal LTM, andWM capacity were estimated with the Wechsler Test of Adult Reading and Peabody Picture Vocabulary Test, delayedrecall of the Hopkins Verbal Learning Test-Revised and Logical Memory of the Wechsler Memory Scale, and Digit SpanTotal, respectively. Intellectual enrichment predicted LTM (B = .54; p = .003) and predicted WM capacity (B = .91;p< .001). WM capacity predicted LTM, (B = .44; p< .001) and fully mediated the relationship between intellectualenrichment (B = .24; p = .27) and LTM (B = .33, p = .03), Sobel test, Z = 3.31, p< .001. These findings implicateWM capacity as an underlying mechanism of cognitive reserve and are an initial first step in understanding the relation-ship between intellectual enrichment, WM, and LTM in MS. (JINS, 2014, 20, 868–872)

Keywords: Multiple sclerosis, Cognitive reserve, Working memory capacity, Long-term memory, Intellectual enrichment,Intelligence

INTRODUCTION

Multiple Sclerosis (MS) is a neurological disease with lesionsand plaques dispersed across the central nervous system,resulting in motor, cognitive, and psychiatric problems.Approximately half of MS patients suffer cognitive impair-ment, including deficits in long-term memory (LTM: 40–65%of patients) leading to reduced quality of life (Chiaravalloti &DeLuca, 2008 for review). There are different reasons whysome MS patients are able to maintain memory functiondespite disease. First, there may be minimal neuropathologyaffecting memory structures in some patients. In addition,recent evidence suggests that MS patients with greater lifetimeintellectual enrichment (frequently estimated with indices of

premorbid verbal intelligence, c.f., Lezak (2004)) are protectedagainst the negative effects of MS neurologic burden onmemory (Sumowski, Wylie, Chiaravalloti, & DeLuca, 2010).This is known as the cognitive reserve hypothesis. Similarfindings have been shown in aging/Alzheimer’s disease(Stern, 2009). A remaining question is how intellectualenrichment affords this cognitive benefit. Recently, Barulli andStern (2013) suggested researchers focus on identifying the“cognitive mechanisms that mediate the relationship betweenbrain challenge and cognitive performance” (p. 507).Working memory (WM) capacity may represent the

cognitive mechanism underlying the protective effects ofintellectual enrichment on LTM (i.e., how cognitive reserveworks). WM is the information processing system involvedin the control, regulation, and maintenance of a limitedamount of information (Miyake & Shah, 1999) and WM istightly linked with LTM. To form new representationsin LTM, information must first pass through WM. WM

Correspondence and reprint requests to: Joshua Sandry, Neuropsychol-ogy and Neuroscience Laboratory, Kessler Foundation, 300 Executive Drive,Suite 70, West Orange, NJ 07052. E-mail: [email protected]

868

system on an alternate measure) as well as improvementsin closely related cognitive domains (moderate transfer todifferent but related cognitive systems, i.e., the secondarymemory [LTM] component of an immediate free recall task).Importantly however, training did not seem to transfer toother distantly related cognitive domains (far transfer, e.g.,intelligence) (Harrison et al., 2013).Given there is evidence for WM training transferring and

improving memory processes in healthy populations, apopulation whose memory performance may arguably befunctioning at or near ceiling levels, it seems reasonable tohypothesize that individuals who are memory impairedwould have even more to gain from training. Past researchdoes support the hypothesis that use of a computerized WMtraining program leads to improvements on measures of WMin MS (Vogt et al., 2009). Unfortunately the Vogt et al. studydid not include a measure of LTM and many training studiessuffer from serious methodological flaws (e.g., lack of ade-quate control groups, interventions that are not theoreticallymotivated), thus, transfer between WM training and LTMremains largely untested in MS and other neurologicalpopulations. The outlook for using WM training as a clinicaltool to improve LTM in MS does appear promising.Future work investigating transfer between cognitive

training and LTM in MS should take special steps to designtraining programs that engage the appropriate underlyingprocesses that contribute to WM. This may increase theprobability that training will transfer between cognitivesystems. These suppositions must be tested and replicated inMS and other neurological populations. Ideal protocols willinclude measures of LTM (delayed recall) to begin to investi-gate improvements and transfer to this memory system.Transfer from cognitive training to LTMmay also result fromtraining other related cognitive systems, for example, trainingprocessing speed or attention. Additionally, many cognitivetraining tasks may overlap and training in one cognitivedomain may result in complimentary training in a relatedcognitive domain. Because of this, it is impossible to assumea training program (or measure), is process-pure (c.f., Jacoby,1991). That is, tasks that train WM may also train processingspeed, attention, or other cognitive systems, and the contrary.Although more females than males are affected by MS, the

present sample was disproportionately female (6:1). Thisshould be kept this in mind when making comparisons across

studies or generalizing these findings to the larger MS popu-lation as a whole. An additional limitation is that the presentstudy relied on transformed values and lacked a control groupfor comparisons. This renders the findings as preliminary.The present study suggests that WM may be the mechanism

through which greater lifetime intellectual enrichment protectsagainst LTM deficits in MS patients. Interventions directedat improving specific sub-processes of the WM system maylead to improvements in cognitive reserve, and concomitantprotection against LTM decline. It remains unclear whetherWMcapacity underlies reserve in other populations, or whetherthis mechanism is specific to reserve in MS patients. Futureresearch is necessary to identify whether WM capacity under-lies reserve in other populations, and whether interventions toimprove WM will lead to improvements in cognitive reserveand LTM.

ACKNOWLEDGMENTS

J.S. supported by National Multiple Sclerosis Society PostdoctoralFellowship Grant MB0024. Study supported by NIH R00 HD060765to J.F.S. The authors have no conflicts of interest to report.

REFERENCES

Barrett, L.F., Tugade, M.M., & Engle, R.W. (2004). Individualdifferences in working memory capacity and dual-processtheories of the mind. Psychological Bulletin, 130(4), 553.

Barulli, D., & Stern, Y. (2013). Efficiency, capacity, compensation,maintenance, plasticity: Emerging concepts in cognitive reserve.Trends in Cognitive Sciences, 17(10), 502–509.

Brandt, J., & Benedict, R.H. (2001). Hopkins Verbal LearningTest, Revised: Professional Manual. Lutz, FL: PsychologicalAssessment Resources.

Chiaravalloti, N.D., & DeLuca, J. (2008). Cognitive impairment inmultiple sclerosis. The Lancet Neurology, 7(12), 1139–1151.

Conway, A.R., Kane, M.J., & Engle, R.W. (2003). Workingmemory capacity and its relation to general intelligence. Trendsin Cognitive Sciences, 7(12), 547–552.

DeLuca, J., Chelune, G.J., Tulsky, D.S., Lengenfelder, J., &Chiaravalloti, N.D. (2004). Is speed of processing or workingmemory the primary information processing deficit in multiplesclerosis? Journal of Clinical and Experimental Neuropsychol-ogy, 26(4), 550–562.

Dunn, L.M., & Dunn, D.M. (2007). Peabody Picture VocabularyTest, Fourth Edition, Manual. Toronto: Pearson PsychCorp.

Fig. 1. Mediation model showing full mediation between Intellectual Enrichment and Long-term Memory. Note. Direct effect ofIntellectual Enrichment (R2 = .13) and indirect effect through the mediator (R2 = .19). Coefficient presented in parentheses indicatesdirect effect. B = unstandardized values. β = standardized values, *p< .01**p< .001.

Mechanisms of cognitive reserve in MS 871

Multiple Sclerosis Journal19(14) 1943 –1946© The Author(s) 2013Reprints and permissions: sagepub.co.uk/journalsPermissions.navDOI: 10.1177/1352458513485980msj.sagepub.com

MULTIPLESCLEROSIS MSJJOURNAL

IntroductionMemory impairment is a prevalent and debilitating symptom of multiple sclerosis (MS),1 but attempts at rehabilitation have been ineffective.2 Retrieval practice (RP) is a powerful memory strategy among healthy persons,3 whereby retention of information is much better when people practice retriev-ing the material (quizzing themselves) rather than restudying the material through massed restudy (“cramming”) or spaced restudy (distributed learning).3,4 Research on RP has been largely limited to healthy college samples, but RP may hold promise as a therapeutic technique for neurologic patients suffering from memory impairment. Indeed, we have recently shown that RP improves memory after a short delay (45 minutes) among memory-impaired patients with MS5 and traumatic brain injury.6 We now investigate whether RP improves memory after both short (30 minutes) and long delays (one week) in memory-impaired MS patients.

MethodsSubject enrollmentTwelve persons with MS7 (Table) and severe memory impair-ment were enrolled (criterion: ≤ second percentile on delayed recall of the Hopkins Verbal Learning Test, Revised8). None of these subjects participated in our previous retrieval practice study,5 making this an independent sample. The institutional review board responsible for the ethical conduct of research at the Kessler Foundation approved this study. Written informed consent was obtained from all subjects.

Experimental procedureIn a within-subjects design, MS patients studied 48 verbal paired associates (Ground-Cold) equally divided across three learning conditions: massed restudy (MR), spaced restudy (SR), and retrieval practice (RP). As illustrated in Figure 1, MR is tantamount to “cramming,” a ubiquitous memory strategy among college students and neurologic patients alike. SR represents distributed learning, recog-nized as superior to MR for more than a century.9 For RP, verbal paired associates were presented on the same sched-ule as SR; however, after the verbal paired associate was presented in its complete form initially (e.g. Ground-Cold), the two subsequent re-exposure trials were framed as cued recall tests (e.g. Ground-_____). A more detailed descrip-tion of learning trials is available in Figure 1 and else-where.5 Dependent measures included delayed recall for half of the verbal paired associates (eight in each condition: MR, SR, and RP) after a short delay (30 minutes), and the

Retrieval practice is a robust memory aid for memory-impaired patients with MS

James F Sumowski1,2, Victoria M Leavitt1,2, Amanda Cohen1, Jessica Paxton1,2, Nancy D Chiaravalloti1,2 and John DeLuca1,2,3

AbstractMemory impairment is prevalent in multiple sclerosis (MS). Retrieval practice is a powerful memory technique whereby retrieving information (quizzing oneself) leads to better memory than restudying. In a within-subjects experiment, 12 memory-impaired MS patients encoded verbal paired associates (VPAs) through massed restudy (MR), spaced restudy (SR), or retrieval practice (RP). Half of VPAs were tested after short delay (30 minutes) and half after long delay (one week). RP robustly improved memory more than restudy. Short delay: MR=15.6%, SR=27.1%, RP=72.9%. Long delay: MR=1.0%, SR=4.2%, RP=24.0%. RP was the best memory technique for nearly all patients after both short and long delays.

KeywordsMultiple sclerosis, memory, neuropsychology, rehabilitation, testing effect

Date received: 29th May 2012; revised: 11th March 2013; accepted: 17th March 2013

1Kessler Foundation Research Center, USA. 2Department of Physical Medicine and Rehabilitation UMDNJ. 3 Department of Neurology and Neurosciences UMDNJ—New Jersey Medical School, USA.

Corresponding author:James F. Sumowski, Neuropsychology and Neuroscience Laboratory, Kessler Foundation Research Center, 300 Executive Drive, Suite 70, West Orange, NJ 07052, USA. Email: [email protected]

485980 MSJ191410.1177/1352458513485980Multiple Sclerosis JournalSumowski et al.2013

Short Report

at Universidad Nacional Aut Mexic on October 7, 2014msj.sagepub.comDownloaded from

43

44

1944 Multiple Sclerosis Journal 19(14)

other half after a long delay (one week). Patients were pre-sented with the first word of each verbal paired associate and were asked to recall the second word.

Statistical analysisRepeated-measures analysis of variance (ANOVA) assessed differences in short delay recall across the three learning

Table. Demographic Data, Neuropsychological Profile and Experimental Memory Outcomes.

Pt Age Edu Dis Dur

HVLT-R DR Raw

HVLT-R DR T-sc

SDMT Raw

PASAT-3 Raw

MFIS Raw

CMDI Raw

CMDI Total T-sc

SD MR

SD SR

SD RP

LD MR

LD SR

LD RP

1 53 16 9 5 23 43 23 54 58 43.4 1 4 3 0 0 02 55 12 7 2 19 34 40 35 68 48.8 2 3 5 0 1 13 42 12 6 5 25 34 56 40 62 45.5 3 3 8 0 0 14 46 16 9 5 25 53 29 9 61 45.0 1 4 8 0 0 15 33 16 14 6 28 50 53 28 56 42.3 0 2 6 0 0 16 57 16 12 2 19 53 57 11 86 58.5 2 2 5 0 0 17 46 16 19 4 19 52 42 68 95 63.3 0 2 6 0 1 28 55 18 26 4 19 33 17 45 77 53.6 1 1 5 0 1 29 60 12 23 5 25 43 37 53 107 69.8 0 0 4 0 0 2

10 47 16 23 6 30 39 45 50 87 59.0 2 1 6 0 1 311 59 12 20 1 19 15 31 49 61 45.0 1 2 7 0 0 412 40 16 20 5 25 22 50 62 79 54.7 2 2 7 1 0 5

Age, education (Edu), and disease duration (Dis Dur) are in years. All patients were women, most had relapsing–remitting multiple sclerosis (RRMS), but two had secondary progressive multiple sclerosis (SPMS) (#s 8 and 9). Pt: patient: HVLT-R DR: Hopkins Verbal Learning Test, Revised Delayed Recall; T-sc: T-scores; SDMT: Symbol Digit Modalities Test, Oral Version; PASAT-3: Paced Auditory Serial Addition Task, 3-second; MFIS: Modified Fatigue Impact Scale; CMDI: Chicago Multidimensional Depression Inventory; SD: short delay of 30 minutes; LD: long delay of one week; MR: massed restudy; SR: spaced restudy; RP: retrieval practice. HVLT-R DR T-scores were derived from normative data published within the HVLT-R manual.8

Figure 1. Sample presentation schedule for VPAs in each learning cognition: MR (white), and SR (light gray), RP (dark gray). Each VPA within the MR condition was presented for three consecutive trials (no intervening trials of other VPAs). Each VPA within the SR or RP conditions was presented in a spaced fashion, with three intervening trials (of other VPAs) between the first presentation and first restudy (or test) trial, and six intervening trials (of other VPAs) between the first and second restudy (or test) trials. This basic presentation schedule was used for all 48 VPAs, totaling 144 six-second trials during a single 14-minute 24-second learning phase (six second slides × 144 slides) viewed on a computer screen. Given that VPAs in each condition were evenly distributed throughout the learning phase, any retroactive or proactive interference or practice effects were also equally distributed across groups. Also, VPAs were counterbalanced across the three different conditions across subjects, as well as across delayed recall tests (30 minutes, one week).VPAs: verbal paired associates; MR: massed restudy; SR: spaced restudy; RP: retrieval practice.


MR= Restudio masivo (mass restudy) 9, 10, 11; 16,17,18... 3 ensayos consecutivos sin interrupciones SR= Restudio espaciado (spaced restudy) 5, 12... · ensayos de otros procedimientos RP= Práctica en recuperación (retrieval practice) 6, 8, 13, 15... 3 ensayos de otros procedimientos

45

MR= Restudio masivo (mass restudy) SR= Restudio espaciado (spaced restudy) RP= Práctica en recuperación (retrieval practice)

Recuerdo demorado (30’)

Recuerdo demorado (Una semana)

Sumowski et al. 1945

conditions: MR, SR, and RP. Next, pairwise comparisons investigated differences in recall across pairs of learning conditions (e.g. MR versus SR). These analyses were repeated for long delay recall.

ResultsThere was a large main effect of learning condition after the short delay (F[2, 22]=60.90, p < 1E−9; η2 = 0.85; Figure 2(a)). Patients recalled 72.9% of verbal paired associates studied through RP, compared to only 15.6% through MR (p < 1E−6) and 27.1% through SR (p < 1E−4). SR led to better memory than MR (p = 0.03). This effect of RP was endur-ing, as the effect of the learning condition remained after the long delay (F = 16.72, p < 1E−4; η2 = 0.60; Figure 2(b)). Patients recalled 24.0% of verbal paired associates studied through RP, compared to only 1.0% through MR (p < 0.001) and 4.2% through SR (p = 0.004). MR and SR did not reliably differ from each other (p = 0.191).

Raw data are provided (Table) so that the benefits of RP can also be examined case by case. Inspection of individual cases revealed that RP was the best memory strategy for nearly every patient at both short and long delays. Moreover, after one week, most patients could not recall a single ver-bal paired associate learned through MR or SR, but nearly all patients recalled at least one verbal paired associate learned through RP, and many recalled at least two or more.

DiscussionRP improved memory much more than restudy strategies in MS patients with severe memory impairment, even after a week-long delay. RP was the most effective memory strat-egy for nearly every patient, bolstering confidence that RP would improve real-life memory functioning. RP is a com-pensatory approach to memory rehabilitation, as it improves memory without repairing/restoring the neurophysiologic basis of memory function. As such, RP will be effective only if patients learn to incorporate RP into their daily rou-tines. For instance, patients wishing to learn information in a newspaper article, training manual, or textbook may engage in intermittent self-quizzing throughout their read-ing (i.e. after each paragraph or page). This act of RP will result in greater subsequent memory than rereading the information multiple times. Despite this, college students4 and MS patients5 mistakenly identify MR (i.e. cramming) as the most effective memory strategy. Therefore, educa-tion, training, and practice will be required for patients to replace MR with the more effective RP technique.

The precise mechanisms underlying the memory bene-fits of RP are unknown, but the effect is consistent with the principle of transfer appropriate processing, which posits that memory is enhanced when there is an overlap between the operations employed during learning and those employed during recall.10 Indeed, given the double

dissociation between brain regions associated with learning and delayed retrieval,11 RP during learning may activate and strengthen the same neural networks during learning that are then used for subsequent delayed recall, thereby facilitating that subsequent recall. Functional neuroimag-ing research is needed to better understand the neurophysi-ologic mechanisms underlying RP.

We acknowledge limitations and future directions. Our sample was relatively small, although this is mitigated somewhat by the within-subjects design and robust results.

Figure 2. Effect of learning condition on the proportion of VPAs correctly recalled after a short delay of 30 minutes (A), and a long delay of one week (B). Error bars represent standard errors.


Estrategias para mejorar la memoria • Mejorar la atención •  Interés por lo que hay que recordar (motivación) • Estrategias internas: asociaciones, mnemotecnias • Estrategias externas: uso de agenda, calendarios,

recordatorios, etc.

46

47


J.J. García-Peñas, et al

Tabla. Efectos adversos (EA) cognitivos y conductuales de los fármacos antiepilépticos de primera, segunda y tercera generación.

Fenobarbital

EA neurotóxicos de clase A en un 61-66%, incluyendo somnolencia, sedación, astenia, apatía, deficiente nivel de alerta y atención, así como pobre coordinación visuomotriz [2,6,8-17,22-26] EA paradójicos o estimulantes en hasta un 50-75%, donde destaca hiperactividad, impulsividad, déficit de atención, trastornos de sueño, agresividad y labilidad emocional [2,6,8-17,22-26] Frecuentes EA neurológicos crónicos con repercusión sobre aprendizaje global, ejecución perceptivomotora, memoria y atención [2,6,8-17,22-26] Potencial implicación en trastornos del neurodesarrollo en hijos de madres epilépticas tratadas con este FAE durante la gestación [20,21]

PrimidonaFrecuentes EA neurotóxicos, incluyendo somnolencia, sedación, astenia, apatía e incoordinación visuomotriz [2,6,8-17,24,26] Frecuencia variable de EA paradójicos como alteraciones de la personalidad, cuadros confusionales y episodios psicóticos [2,6,8-17,24,26]Frecuentes EA neurológicos evolutivos con repercusión sobre la memoria y la atención [2,6,8-17,24,26]

Fenitoína

Frecuente neurotoxicidad de clase A, con somnolencia, confusión y descompensación paradójica de la epilepsia (encefalopatía aguda por fenitoína) [2,6,8-17,24,27] Potencial reacción paradójica con hiperactividad e impulsividad en niños con encefalopatías crónicas [2,6,8-17,24,27]Frecuente deterioro neurocognitivo evolutivo con alteración del aprendizaje, la atención, la capacidad de concentración, la memoria operativa y la velocidad de procesamiento [2,6,8-17,24,27]. Potencial implicación en trastornos del neurodesarrollo en hijos de madres epilépticas tratadas con este FAE durante la gestación [20,21]

EtosuximidaInfrecuentes EA neurológicos tipo A como somnolencia, lentitud de pensamiento, síndrome confusional o psicosis aguda [2,6,8-17,28]Infrecuentes EA neurológicos crónicos como trastornos de aprendizaje, déficit de atención, alteraciones de la memoria y el lenguaje, depresión, cambios de personalidad y alteraciones del comportamiento [2,28]

CarbamacepinaEA neurológicos de clase A en un 30-40% de los casos, incluyendo sedación, somnolencia, ansiedad e incoordinación visuomotriz [2,6,8-17,24,31] Infrecuentes EA neurológicos crónicos como déficit de atención, hiperactividad, trastornos de conducta o problemas de aprendizaje [2,6,8-17,24,31]Riesgo de encefalopatía epiléptica evolutiva al tratar epilepsias rolándicas benignas [8]

Valproato sódico

Frecuencia variable de EA neurológicos de clase A como somnolencia, irritabilidad, alteración del sueño o incoordinación motriz [2,6,8-17,24,28-30] Infrecuente encefalopatía hiperamoniémica con afectación progresiva del nivel de conciencia, principalmente en politerapia con fenobarbital o topiramato [2,6,8-17,29,30] EA neurológicos crónicos en un 15-20% de los casos con clínica de déficit de atención, hiperactividad, impulsividad o trastornos del sueño [2,6,8-17,23,24,28] Se han descrito casos de retraso madurativo global y signos de trastorno del espectro autista como expresión de afectación neurológica en hijos de madres gestantes tratadas con VPA [20,21]

Benzodiacepinas, clonacepam, clobazam

Alta tasa de EA neurológicos de clase A, como somnolencia, lentitud mental e incoordinación visuomotriz, sobre todo con clonacepam [2,6,8-17]Potenciales EA paradójicos estimulantes como hipercinesia, impulsividad, insomnio y delirio [2,6,8-17]Frecuente deterioro crónico cognitivo sobre atención, memoria y capacidad de concentración [2,6,8-17]

Vigabatrina

Frecuentes EA neurológicos de clase A como somnolencia, sedación y ansiedad [2,8,10,12,14-18,31] Potencial riesgo de reacciones paradójicas estimulantes como impulsividad, hipercinesia, agitación psicomotriz, delirio e insomnio [2,8,10,12,14-18,31]No produce EA crónicos sobre las funciones intelectuales [31] Potencial riesgo evolutivo de alteraciones conductuales en un 1-4% [2,8,10,12,14-18,31]

Lamotrigina

Baja incidencia de EA de clase A como somnolencia, astenia o insomnio [2,8,10,12,14-18,32] Infrecuentes reacciones paradójicas como agitación psicomotriz [2,8,10,12,14-18,32] No produce EA crónicos sobre las funciones intelectuales [32]Efecto beneficioso neurocognitivo y conductual crónico al mejorar el nivel de alerta y atención [2,8,10,18,32] Potencial riesgo de alteraciones comportamentales evolutivas con agresividad y conducta oposicionista-desafiante [2,8,10,12,14-18,32]

FelbamatoBaja frecuencia de EA de clase A como somnolencia, ansiedad, trastornos del lenguaje, irritabilidad e insomnio [10,18,33] Ocasionales EA crónicos como problemas de concentración, déficit de atención y lentitud de respuestas [10,18,33]

GabapentinaFrecuentes EA neurotóxicos de clase A como somnolencia, sedación, mareos e incoordinación visuomotriz [2,10,18,34]Potencial riesgo de reacciones paradójicas estimulantes como irritabilidad, hipercinesia, agitación, agresividad e insomnio [2,10,18,34]No produce EA crónicos sobre las funciones intelectuales [2,34]

Tiagabina Los EA neurotóxicos aparecen en un 30-40% de los casos e incluyen somnolencia, astenia, nerviosismo, déficit de atención y trastorno del lenguaje [2,10,18,35] Potencial riesgo de EA crónicos como alteraciones conductuales y psicosis en hasta un 0,8% de los tratados [2,10,18,35]

PregabalinaFrecuentes EA de clase A como somnolencia e incoordinación visuomotriz [10,36]No parece producir EA crónicos sobre las funciones intelectuales [36]

48


J.J. García-Peñas, et al

Tabla. Efectos adversos (EA) cognitivos y conductuales de los fármacos antiepilépticos de primera, segunda y tercera generación.

Fenobarbital

EA neurotóxicos de clase A en un 61-66%, incluyendo somnolencia, sedación, astenia, apatía, deficiente nivel de alerta y atención, así como pobre coordinación visuomotriz [2,6,8-17,22-26] EA paradójicos o estimulantes en hasta un 50-75%, donde destaca hiperactividad, impulsividad, déficit de atención, trastornos de sueño, agresividad y labilidad emocional [2,6,8-17,22-26] Frecuentes EA neurológicos crónicos con repercusión sobre aprendizaje global, ejecución perceptivomotora, memoria y atención [2,6,8-17,22-26] Potencial implicación en trastornos del neurodesarrollo en hijos de madres epilépticas tratadas con este FAE durante la gestación [20,21]

PrimidonaFrecuentes EA neurotóxicos, incluyendo somnolencia, sedación, astenia, apatía e incoordinación visuomotriz [2,6,8-17,24,26] Frecuencia variable de EA paradójicos como alteraciones de la personalidad, cuadros confusionales y episodios psicóticos [2,6,8-17,24,26]Frecuentes EA neurológicos evolutivos con repercusión sobre la memoria y la atención [2,6,8-17,24,26]

Fenitoína

Frecuente neurotoxicidad de clase A, con somnolencia, confusión y descompensación paradójica de la epilepsia (encefalopatía aguda por fenitoína) [2,6,8-17,24,27] Potencial reacción paradójica con hiperactividad e impulsividad en niños con encefalopatías crónicas [2,6,8-17,24,27]Frecuente deterioro neurocognitivo evolutivo con alteración del aprendizaje, la atención, la capacidad de concentración, la memoria operativa y la velocidad de procesamiento [2,6,8-17,24,27]. Potencial implicación en trastornos del neurodesarrollo en hijos de madres epilépticas tratadas con este FAE durante la gestación [20,21]

EtosuximidaInfrecuentes EA neurológicos tipo A como somnolencia, lentitud de pensamiento, síndrome confusional o psicosis aguda [2,6,8-17,28]Infrecuentes EA neurológicos crónicos como trastornos de aprendizaje, déficit de atención, alteraciones de la memoria y el lenguaje, depresión, cambios de personalidad y alteraciones del comportamiento [2,28]

CarbamacepinaEA neurológicos de clase A en un 30-40% de los casos, incluyendo sedación, somnolencia, ansiedad e incoordinación visuomotriz [2,6,8-17,24,31] Infrecuentes EA neurológicos crónicos como déficit de atención, hiperactividad, trastornos de conducta o problemas de aprendizaje [2,6,8-17,24,31]Riesgo de encefalopatía epiléptica evolutiva al tratar epilepsias rolándicas benignas [8]

Valproato sódico

Frecuencia variable de EA neurológicos de clase A como somnolencia, irritabilidad, alteración del sueño o incoordinación motriz [2,6,8-17,24,28-30] Infrecuente encefalopatía hiperamoniémica con afectación progresiva del nivel de conciencia, principalmente en politerapia con fenobarbital o topiramato [2,6,8-17,29,30] EA neurológicos crónicos en un 15-20% de los casos con clínica de déficit de atención, hiperactividad, impulsividad o trastornos del sueño [2,6,8-17,23,24,28] Se han descrito casos de retraso madurativo global y signos de trastorno del espectro autista como expresión de afectación neurológica en hijos de madres gestantes tratadas con VPA [20,21]

Benzodiacepinas, clonacepam, clobazam

Alta tasa de EA neurológicos de clase A, como somnolencia, lentitud mental e incoordinación visuomotriz, sobre todo con clonacepam [2,6,8-17]Potenciales EA paradójicos estimulantes como hipercinesia, impulsividad, insomnio y delirio [2,6,8-17]Frecuente deterioro crónico cognitivo sobre atención, memoria y capacidad de concentración [2,6,8-17]

Vigabatrina

Frecuentes EA neurológicos de clase A como somnolencia, sedación y ansiedad [2,8,10,12,14-18,31] Potencial riesgo de reacciones paradójicas estimulantes como impulsividad, hipercinesia, agitación psicomotriz, delirio e insomnio [2,8,10,12,14-18,31]No produce EA crónicos sobre las funciones intelectuales [31] Potencial riesgo evolutivo de alteraciones conductuales en un 1-4% [2,8,10,12,14-18,31]

Lamotrigina

Baja incidencia de EA de clase A como somnolencia, astenia o insomnio [2,8,10,12,14-18,32] Infrecuentes reacciones paradójicas como agitación psicomotriz [2,8,10,12,14-18,32] No produce EA crónicos sobre las funciones intelectuales [32]Efecto beneficioso neurocognitivo y conductual crónico al mejorar el nivel de alerta y atención [2,8,10,18,32] Potencial riesgo de alteraciones comportamentales evolutivas con agresividad y conducta oposicionista-desafiante [2,8,10,12,14-18,32]

FelbamatoBaja frecuencia de EA de clase A como somnolencia, ansiedad, trastornos del lenguaje, irritabilidad e insomnio [10,18,33] Ocasionales EA crónicos como problemas de concentración, déficit de atención y lentitud de respuestas [10,18,33]

GabapentinaFrecuentes EA neurotóxicos de clase A como somnolencia, sedación, mareos e incoordinación visuomotriz [2,10,18,34]Potencial riesgo de reacciones paradójicas estimulantes como irritabilidad, hipercinesia, agitación, agresividad e insomnio [2,10,18,34]No produce EA crónicos sobre las funciones intelectuales [2,34]

Tiagabina Los EA neurotóxicos aparecen en un 30-40% de los casos e incluyen somnolencia, astenia, nerviosismo, déficit de atención y trastorno del lenguaje [2,10,18,35] Potencial riesgo de EA crónicos como alteraciones conductuales y psicosis en hasta un 0,8% de los tratados [2,10,18,35]

PregabalinaFrecuentes EA de clase A como somnolencia e incoordinación visuomotriz [10,36]No parece producir EA crónicos sobre las funciones intelectuales [36]

49

S39www.neurologia.com Rev Neurol 2014; 58 (Supl 1): S37-S42

Epilepsia. Neurofisiología y trastornos del desarrollo

poca información disponible y ésta es muy contra-dictoria [2,7,10,18]. Sin embargo, se conocen bien los efectos adversos cognitivos específicos de algu-nos FAE como el topiramato sobre la memoria, la atención y el lenguaje, con un nivel de evidencia clase III [7,8,10,18].

El objetivo de esta revisión es conocer el papel de los efectos adversos de los FAE sobre el aprendi-zaje y la conducta en el niño epiléptico, analizando los aspectos globales del tratamiento y las peculia-ridades individuales de los distintos FAE de prime-ra, segunda y tercera generación.

Papel de los efectos adversos de los FAE sobre la cognición, el aprendizaje y la conducta

Los efectos adversos de los FAE sobre el aprendi-zaje, la conducta y las funciones cognitivas pueden presentarse de cuatro formas diferentes, depen-diendo de su relación con el mecanismo de acción del FAE, la dosis administrada, los niveles séricos,

la respuesta biológica del paciente y la aparición precoz, crónica o diferida del efecto adverso noci-vo [19]:– Efectos adversos de clase A o efectos adversos neu-

rotóxicos directos. Son aquellos que dependen del mecanismo de acción del FAE y guardan una relación directa con la dosis administrada o con los niveles plasmáticos del fármaco. Por ello, son efectos adversos previsibles y anticipables. Son efectos frecuentes, pero no revisten gravedad en la mayoría de los casos. Suelen aparecer desde las primeras fases del tratamiento y pueden mi-nimizarse ralentizando la escalada de dosis del FAE. Con frecuencia, se llega a establecer una tolerancia parcial o total a estos efectos adversos con el transcurso del tiempo. Los síntomas más frecuentes de las reacciones tipo A son la som-nolencia y el enlentecimiento del curso del pen-samiento [2,7-19].

– Efectos adversos de clase B o reacciones adversas idiosincrásicas. Son aquellos que no dependen únicamente de las características farmacológicas

OxcarbacepinaFrecuentes EA de clase A como somnolencia e incoordinación visuomotriz [2,8,10,12,14-18,27,37] No hay evidencia de deterioro cognitivo crónico por oxcarbacepina [27,37]

Topiramato

Frecuentes EA de clase A como somnolencia, lentitud mental, nerviosismo, ansiedad y alteraciones de conducta [2,8,10,12,14-18,27,38]Potencial riesgo de encefalopatía hiperamoniémica en pacientes tratados con la asociación de valproato sódico y topiramato [30,38] Frecuencia variable de EA neurológicos crónicos, principalmente en politerapia, con alteración de atención y memoria verbal, donde la anomia es el EA más característico de este cuadro [2,8,10,12,14-18,27,38]

LevetiracetamFrecuentes EA de clase A como somnolencia, astenia, inquietud, irritabilidad e insomnio [2,8,10,12,14-18,27,39] Efecto beneficioso neurocognitivo crónico al mejorar el nivel de alerta y atención [2,10,39] Potencial riesgo de alteraciones conductuales evolutivas como hiperactividad, impulsividad, agresividad e, incluso, algún caso de psicosis [2,8,10,12,14-18,27,39]

ZonisamidaFrecuentes EA de clase A como somnolencia, astenia, agitación, irritabilidad e insomnio [2,10,40] Potencial afectación neurocognitiva crónica que incide sobre la atención, la capacidad de concentración y la memoria verbal [2,10,40]Potencial riesgo de alteraciones del comportamiento graves evolutivas [2,10,40]

RufinamidaFrecuentes EA de clase A como somnolencia, irritabilidad e incoordinación motriz, principalmente en politerapia con valproato sódico [10,41] No existe evidencia de potencial deterioro cognitivo crónico con rufinamida [41]

LacosamidaFrecuentes EA de clase A que afectan la coordinación visuomotriz [10,42] Perfil cognitivo crónico seguro [42]

EstiripentolHay EA de clase A en un 50-75% de los casos y se relacionan principalmente con la potenciación de otros FAE como el valproato sódico y el clobazam; destaca somnolencia, incoordinación visuomotriz, irritabilidad e insomnio [43]

Acetato de eslicarbacepina

Frecuentes EA de clase A como somnolencia e incoordinación visuomotriz [10,44]No se describe potencial deterioro neurocognitivo evolutivo [44]

Perampanel Frecuentes EA de clase A, destaca somnolencia e incoordinación visuomotriz [45]

Tabla. Efectos adversos (EA) cognitivos y conductuales de los fármacos antiepilépticos de primera, segunda y tercera generación (cont.).

principales sÍndromes neurolÓgicos: … · dose en volumen según transcurrían los años en ......

Documents