10/31/2011

1

L37. SPIKE TIMING DEPENDENT PLASTICITY: Hebbian and Anti‐Hebbian plasticity of 

SynapsesSynapses

October 31, 2011

C. D. Hopkins

In the previous lecture we learned how Mormyrid electric fish produce an ‘electromotor’ command and then receive an electrosensory response after a delay. The motor output and the sensory response are time-locked to each other. The fish builds an expectation of the sensory response so that it can remove it from the rest of the sensory inflow, not by inhibitory blanking, but by forming a negative image of the expected input.

2Meek, Grant and Bell. The Journal of Experimental Biology 202, 1291–1300 (1999)

bulbar command associated n.midbrain command associated n.juxtalemniscal cells – to NELL

Principal Cell PlasticityEOD command Recording from a principal cell in the ELL of

mormyrid shows no response to the EOD command alone.

When an EOD substitute (artificial electrical stimulus) stimulus is turned on, in synchrony with the EOD command, the ELL cell responds – first with a pause in firing, then a burst of spikes. The raster plot shows a decrease in spike density followed by an increase.

After 9 minutes of pairing, however, the pause becomes “filled in” while the burst is weaker.

3

Turning the stimulus off, now one finds that the command alone produces a burst-pause, exactly the negative image of the sensory response to the electrical stimulus.

After 9 minutes of command disharges alone with no stimulation, the burst/pause weakens and comes back to spontaneous activity.

This is an example of a modifiable efference copy.

In the previous lecture we also looked at a case of negative image formation in an electric fish from South America, where

there is no corollary discharge for the EOD command.

4

Apteronotus leptorhynchus (a wave species with a nerve derived electric organ)

Eigenmannia virescens, a wave species with a modified muscle derived electric organ

Bastian: Sensory Consequence of Movement of Tail

Experimental paradigm:

1) Tail movement controlled by motor.

2) Electric discharge continues d ti l t

5

and stimulates electroreceptors.

3) Electrode near skin monitors the EOD amplitude

JOSEPH BASTIAN (1999) PLASTICITY OF FEEDBACK INPUTS IN THE APTERONOTID ELECTROSENSORYSYSTEM J. Experimental Biology 202:1327-1337

Electrosensory Feedback From Tail Bending is Cancelled in ELL

6

The cancellation is modifiable over time.

JOSEPH BASTIAN (1999) PLASTICITY OF FEEDBACK INPUTS IN THE APTERONOTID ELECTROSENSORY SYSTEM J. Experimental Biology 202:1327-1337

10/31/2011

2

If tail is not bent during artificial Amplitude modulation, there is no effect

7

The same experiment was done with Eigenmannia, but this time without the benefit of propioceptive input (no tail bending)

B1) The whole body is stimulated with an amplitude modulated EOD.

experiments were done on Eigenmannia

8

B2) Whole body plus a local AM

B3) After 10 mintues, however the effect of the local AM goes away.

B4) Whole body AM alone, now the whole body AM forms a negative image

B5) after 10 minutes of AM alone,, local AM off.

Pairing of electrosensory stimulation with ventilation movements in Skate also produces a negative image of expected sensory input

The principal cell in the Dorsal OctavolateralNucleus (DON) of skate.

No response to expiration, inspiration movements of gill.

When ventilation is paired with electrosensorystimulus, the principal cell responds

9

Montgomery and Bodznick, 1994

After 25 minutes, weaker response to stimulus.

Now, when the stimulus goes off, the ventilation response produces a negative image

10

In principal cells of the mormyrid ELL, pairing a EOD command discharge with an intracellular stimulus can result in formation of negative image (only under strong stimulation)

11Bell, C. C., Caputi, A., Grant, K. and Serrier, J. (1993). Storage

of a sensory pattern by anti-hebbian synaptic plasticity in an electric fish. Proc Natl Acad Sci U S A 90, 4650-4654.

The principle cell in response to EODc + weak intracellular, no plasticityIn response to strong intracellular stimulus which evokes a broad spike, plasticity occurs.

The Basic Architecture

1. Bell, C. C., Han, V. and Sawtell, N. B. (2008). Cerebellum-like structures and their implications for cerebellar function. AnnuRev Neurosci 31, 1-24.

10/31/2011

3

13

Theory to Explain Negative Image

Sensory input arrives from below and stimulates pyramidal cell to fire.

Predictive input comes from above

14

through parallel fibers.

Synaptic modification changes strength of predictive input as a consequence of simultaneous action.

Recording Plasticity in Slice Preparation

Recording intracellularly from an ELL cell from mormyromast region of mormyrid ELL.

15

Stimulate parallel fibers with two sets of electrodes in molecular layer of ELL.

Spike timing dependent plasticity is anti-hebbian. If the broad spike occurs after the EPSP, the synaptic weight is decreased. If the broad spike occurs before the EPSP, it strengthens. This sculpts a negative image of the expected input.

causalacausal

Percentage change in excitatory postsynaptic potential (EPSP) amplitude plotted against the delay between EPSP onset and the broad spike peak during pairing. A negative delay with regard to the previous spike and a positive delay with regard to the following spike were present for each pairing. The shorter of these two delays is plotted. Filled circles, significant changes; open circles, non-significant changes (at P<0.01).

animation rule

T (post-pre)

Bell CC, Han VZ, Sugawara Y, Grant K. 1997. Synaptic plasticity in a cerebellum-like structure depends on temporal order. Nature 387:278–81

17

presynapticspike before post

post synaptic spike before pre

Note: the sign of the time axis here is reversed compared to previous slide

18

Defining Spike-Timing-Dependent Plasticity (A) A presynaptic cell connected to a postsynaptic cell repeatedly spiking just before the latter is in part causing it to spike, while the opposite order is acausal. (B) In typical STDP, causal activity results in long-term potentiation (LTP), while acausal activity elicits long-term depression (LTD; Markram et al., 1997b; Bi and Poo, 1998; Zhang et al., 1998). At somecortical synapses, the temporal window for LTD (dashed gray line) is extended (Feldman, 2000; Sjöström et al., 2001). These temporal windows are often also activity dependent, with LTP being absent at low-frequency(gray continuous line, Markram et al., 1997b; Sjöström et al., 2001), and postsynaptic bursting relaxing the LTD timing requirements to hundreds of milliseconds (Debanne et al., 1994; Sjöström et al., 2003).

T(pre-post)

Henry Markram1*,Wulfram Gerstner1 and Per Jesper Sjöström2,3 Frontiers in Synaptic Neuroscience. August 2011

10/31/2011

4

Donald Hebb“Let us assume that the persistence or repetition of areverberatory activity (or “trace”) tends to induce lasting cellularchanges that add to its stability.[. . .] When an axon of cell A is nearenough to excite a cell B and repeatedly or persistently takes part infiring it, some growth process or metabolic change takes place in oneor both cells such that A’s efficiency, as one of the cells firing B, isincreased.” (Hebb, 1949).

Henry Markram1*,Wulfram Gerstner1 and Per Jesper Sjöström2,3 Frontiers in Synaptic Neuroscience. August 2011

normal STDP (Hebbian)

reversed STDP(antiHebbian)

20Natalia Caporale and Yang Dan (2008)

Methods (continued)Identification of Mossy Fibers in Egp.

1) Previous studies by Bell (1991) recorded from cells in EGp with similar properties and found * no synaptic activity, *spikes arising directly from baseline.

2) Several fills with biocytin (3 from proprioceptive input, 2 with EODc input)

3) No neurons responded to both EODc and propioception.

4) Pairing of EODCD synaptic activity and narrow spike activity by averaging responses before and after EODCD, narrow spikes removed.

Cerebellum has always attracted neuroscientists interested in how the brain works. The orderly structure, the size, the individuality of the cellular players, the complexity.

Ramón y Cajal 1894

The main players in cerebellar function are:Granule cells: which receive input from mossy fibers, and send their

outputs to Purkinje cells via parallel fibersPurkinje cells: the “Principal Cell” of the CerebellumClimbing fibers: bring “error signals” from the inferior olive to ONE

Purkinje cell.

Brains of vertebrates color coded by brain area. Cerebellum in orange

10/31/2011

5

Cerebellum

Marr, D. (1969). A theory of cerebellar cortex. J. Physiol. 202, 437–470.

Rationale

Albus, James A. 1971). A theory of cerebellar function. Math. Biosci. 10, 25–61.

“Influential theories of cerebellar function have posited that highly selective or sparse coding in Granule Cells allows Purkinje cells to acquire selective responses through associative synaptic plasticity” “Similarly for cerebellum-like circuits, such as the electrosensory lobe (ELL) of mormyrid fish, Granule Cells that selectively encode specific combinations of sensory and/or motor signals could allow Purkinje-like cells to generate more specific negative images.”“Nothing is yet known about how such signals are recoded in GCs or about the significance of GC input representations for the generation of negative images.”

Overall theme for cerebellar function. A direct pathway from input to output (i.e. the vestibular system to the eye muscles in the case of the VOR) is accompanied by an “indirect pathway” through the cerebellum – vestibular collateral (mossy fiber) to granule cell to parallel fiber to Purkinje cell. Parallel fibers excite Purkinje cells on synaptic spines. When there is an error in the VOR, information about image slip on the retina comes from the inferior. This is a strong depolarization which causes a dendritic spike in the Purkinje cell, and this causes synaptic weights from parallel fibers to depress, weakening the inhibitory input on the vestibular neurons, thereby adjusting the gain of the eye reflex.

Cerebellum and “cerebellar-like” structures

Most vertebrates possess both a cerebellum and other brain structures with cerebellum-like architectures.

1) A MOLECULAR LAYER with many parallel fibers + dendrites of Purkinje-like cells.

2) A LARGE NUMBER OF GRANULE CELLS: which carry information from many central structures about the state of the

molec

deep

central structures about the state of the animal: sensory information from a large number of senses, corollary discharges from motor centers (predictive of sensory input to the principle cells).

3) PRINCIPAL CELLS with SPINES. Parallel fibers terminate on spines of principal cells; or on the smooth dendrites of inhibitory stellate cells.

Cerebellum-like structures sometimes serve as adaptive sensory filters that predict sensory inputs into the deep layers using the associated parallel fiber inputs in the molecular layer.

10/31/2011

6

The cerebellar-like structure is defined by the molecular layer, the principal cell, and the numerous inputs from granule cells via parallel fibers. structure.

The Basic Architecture

1. Bell, C. C., Han, V. and Sawtell, N. B. (2008). Cerebellum-like structures and their implications for cerebellar function. Annu Rev N i 31 1 24

34

Gnathonemus petersii

36

10/31/2011

7

37

Anatomy

Injection of biotinylated dextran into molecular layer of ELL labels:MG cells (gabaergic cells in ELL) granule cells which project via parallel fibers to MGcell dendrites.

Many Mossy Fibers provide propioceptive sensory information from tail.

(most common type of unit)

responds to tail bend, not EODc.

Spike rate unchanged by EOD.

Firing rate is 20-150 Hz.

Has a best tail angle

most preferred extremes. A few intermediate

lg

Spike rate encodes tail angle shown for 3 units that prefer contralateral bend, 3 that prefer ipsilateral bend.

Some proprioceptor units respond to schnauzenorgandisplacement, trunk displacement or fin.

angles

Some mossy fibers respond to EOD command

Putative MossyFiber cell responding to EODc.

No spontaneous activity.

Response precedes EOD. occasion

3 units showing differing latencies of responses. Stereotyped rasters.

Summed histograms of spikes per command from 30 such units.

Active for up to 400 msecafter EODc

occasional cell had burst, then long latency burst.

Other Examples of Removing Expected Inputs

• Adaptation of receptors or neurons to maintained stimuli removes responses to constant stimulus.

L t l i hibiti t d l l

41

• Lateral inhibition: remove expected mean levels over space.”

Both methods use simple non-plastic cellular changes (self inhibition, intra-cellular inhibition).

1: Bastian J. Pyramidal-cell plasticity in weakly electric fish: a mechanism for attenuating responses to reafferent electrosensory inputs. J Comp Physiol A. 1995Jan;176(1):63-73. PubMed PMID: 7823309.

2: Bastian J. Plasticity in an electrosensory system. I. General features of adynamic sensory filter. J Neurophysiol. 1996 Oct;76(4):2483-96. PubMed PMID:8899621.

3: Bastian J. Plasticity of feedback inputs in the apteronotid electrosensorysystem. J Exp Biol. 1999 May;202(Pt 10):1327-37. Review. PubMed PMID: 10210673.