Music increases frontal EEG coherence during verbal

learning

David A. Peterson a,b,c, , Michael H. Thaut b∗ ,ca Department of Computer Science, Colorado State Universityb Program in Molecular, Cellular, and Integrative Neuroscience Colorado State Universityc Center for Biomedical Research in Music, Colorado State University

Ranelle Johnson

Neuroscience Letters 412 (2007) 217-221

Introduction (others said…)• Memory may be subserved by oscillations

in recurrent networks within and between brain regions (in theory)

• Increased multi band spectral power in the EEG during encoding is associated with successful subsequent word recall

Introduction• In an earlier study…

– verbal learning is associated with broadband increases in EEG power spectra

– Music influences the topographic distribution of the increased spectral power

• Present study…– examining spatial coherence in EEG

measured during learning phase

Independent Variable

• Treatment Groups– Learning to recall in conventional speaking

voice– Learning to recall while singing to melody

Dependent Variable

Theoretical Construct- verbal learning

Operational Definition- Transition from not being able to recall to being able to recall a word that is repeatedly presented in the AVLT

Dependent Variable

• Theoretical Construct– Learning related changes in coherence

(LRCC)• Operational Definition

– Percent increase or decrease in coherence comparing “first recalled” words to the same words not recalled during the immediately preceding trial (all pairs of not learned/learned words)

Hypothesis

• Learning that persists over short- and long- delays will be associated with “learning related change in coherence” (LRCC) in frontal EEG

Hypothesis

• The temporally structured learning template provided by music will strengthen LRCC patterns in frontal EEG compared to conventional spoken learning.

Subjects

• 16 healthy right-handed volunteers– Normal hearing– No history- neurological or psychiatric conditions

• Randomly assigned to one of two experimental conditions in a in between-subjects design

• Age range– 18-26 (mean=19.8, SD=2.8)– 18-21 (mean=19.0, SD=1.0)

• Each group contained 7 females

Method

• Rey’s Auditory Verbal Learning Test (AVLT)– 15 semantically unrelated words – Repeated in 5 learning trials– Subjects free recalled as many words as

possible after each recall– 6th trial- distracter list, 20 minute visual

Method

Fig. 1. Rey’s Auditory Verbal Learning Test (AVLT) and the operational definitions of:• learning (thickest arrows, during the learning trials—e.g. words 2, 14, 15);• short-delay memory (medium thick arrows to M1—e.g. words 2, 15);• long-delay memory (thinnest arrows to M2—e.g. word 2).

Method

• Pre-recorded female voice (both conditions)

• Music condition- sang simple, repetitive and unfamiliar melody

• Made both groups’ list of words same durations (15sec)

Method

• Electroencephalogram (EEG)– Continuous EEG from continuous 32 scalp

electrodes – Neuroscan’s QuickCap using low- and high-

frequencies 1 and 100 Hz, 1kHz sampling frequency

– Computed interelectrode coherence over 500 ms window 250ms after each word’s onset

• For electrode pairs in theta, alpha, gamma frequency bands

Analysis

• Computed coherence for– left and right prefrontal areas, within and

between• LRCC for each group- compared to no

change using one-tailed t-test, alpha=.05• LRCC between the two groups- compared

using two-tailed t-test, alpha=.05• Bonferroni adjustment by factor of 3 for

multiple comparison

Results (performance)

• Both groups recalled about 6 more words on the last learned trial than on the first– mean=11 & 4.9 (spoken) – mean=9.7 & 4.3 (music)

• Significant improvement in performance– t(16)= 9.6 & 6.3, p<0.0001

• Recall was not significantly different between spoken and musical groups on any trial – t(16)<1.4, p>0.1

Results (spoken)

• Involves a mix of (+) & (-) LRCC• None of LRCC values differed significantly

from zero• Except…

– Negative right frontal gamma LRCC– t(42)=2.2, p=.03 and t(36)= 2.2, p=.03

Results (musical)

• Involved (+) LRCC within & between the hemispheres

• Short-delays– Increased frontal coherence significant for left

gamma, t(39)=2.6, p=.003– Interhemispheric theta, t(39)=2.6, p=.01

Results (musical)• Between Group Comparisons

– Higher LRCC for music group in all 3 frequency bands– Music showed greater increase in theta coherence for

short- & long- delay learning • t(81)=2.1, p=.04 (short- delay)

– Music showed increase and spoken a decrease for right alpha coherence (both short- and long- delay)

• t(66)=2.5, p=.01 (long delay learning)– Music showed greater increase (spoken decrease) in

right gamma coherence in b/t groups for long- delay learning

• t(66)=2.7, p=.009

Results

In each cell: .Values are mean coherence relative to previous unlearned trial (i.e. 100 is no change in coherence). Boldvalues indicates p < 0.05 in one-tailed T-test after Bonferroni correction. Highlighted values indicates p < 0.05 in between-group, two-tailed T-test after Bonferronicorrection.

Results

Fig. 2. Learning-related change in coherence (LRCC). Left: LRCC for short-delay recall; right: LRCC for long-delay recall. Scale bar is percent change. Straightline and box overlays indicate absolute changes in LRCC greater than 5%: box for local quadrant LRCC, and line for interquadrant (interhemispheric) LRCC. Thinbox (e.g. spoken group’s long-delay right frontal gamma LRCC) indicates a decrease of greater than 5%.

Discussion

• Music condition had increased frontal coherence whereas spoken condition had no significant change

• Music group had stronger temporal synchronization in frontal areas

Discussion

• Lack of coherence in spoken condition may be due to form of measure

• Spoken learning involves more focal changes

• Musical learning shows more topographical broader network synchronization

Discussion

• Performance effect “nullified”• Transfer appropriate processing theory

– Subjects asked to recall material in different way that it was encoded

• Physiological results not due to differences in performance

• Physiological results not due to different sensory processing (music vs. spoken stimuli)– Data for LRCC measured with recalled word

compared to the same word not recalled in previous trial

Discussion

• How does music effect synchronization then?– Early attentional mechanism

• Selective attention associated with greater coherence with multiple spatial skills

– Music is known to form expectancy, listeners can predict musical aspects, this could increase coherence

– Studies suggest that music related processing involves more widely distributed subcortical and cortical networks

What do I think?

• Uuuuuummmmm???• I don’t know enough about the interpretations of

EEG to really be able to criticize a whole lot…• But like they said, use more subjects• They could try testing performance by having

the words sung back and see if it makes a difference on performance

• I don’t understand how you can measure a not recalled word

BUT MOST IMPORTANTLY….

– widely distributed

z, t, F =