Limits...
Theta coordinated error-driven learning in the hippocampus.

Ketz N, Morkonda SG, O'Reilly RC - PLoS Comput. Biol. (2013)

Bottom Line: We show that the differential phase relationships of hippocampal subfields within the overall theta rhythm enable a powerful form of error-driven learning, which results in significantly greater capacity, as shown in computer simulations.In one phase of the theta cycle, the bidirectional connectivity between CA1 and entorhinal cortex can be trained in an error-driven fashion to learn to effectively encode the cortical inputs in a compact and sparse form over CA1.Taken together, these new learning dynamics enable a much more robust, high-capacity model of hippocampal learning than was available previously under the classical Hebbian model.

View Article: PubMed Central - PubMed

Affiliation: Department of Psychology, University of Colorado Boulder, Boulder, Colorado, United States of America.

ABSTRACT
The learning mechanism in the hippocampus has almost universally been assumed to be Hebbian in nature, where individual neurons in an engram join together with synaptic weight increases to support facilitated recall of memories later. However, it is also widely known that Hebbian learning mechanisms impose significant capacity constraints, and are generally less computationally powerful than learning mechanisms that take advantage of error signals. We show that the differential phase relationships of hippocampal subfields within the overall theta rhythm enable a powerful form of error-driven learning, which results in significantly greater capacity, as shown in computer simulations. In one phase of the theta cycle, the bidirectional connectivity between CA1 and entorhinal cortex can be trained in an error-driven fashion to learn to effectively encode the cortical inputs in a compact and sparse form over CA1. In a subsequent portion of the theta cycle, the system attempts to recall an existing memory, via the pathway from entorhinal cortex to CA3 and CA1. Finally the full theta cycle completes when a strong target encoding representation of the current input is imposed onto the CA1 via direct projections from entorhinal cortex. The difference between this target encoding and the attempted recall of the same representation on CA1 constitutes an error signal that can drive the learning of CA3 to CA1 synapses. This CA3 to CA1 pathway is critical for enabling full reinstatement of recalled hippocampal memories out in cortex. Taken together, these new learning dynamics enable a much more robust, high-capacity model of hippocampal learning than was available previously under the classical Hebbian model.

Show MeSH
Diagram of relation to physiological data and computational model.Figure adapted from [6]A) Relation of current-source/sink analysis within subfields of the hippocampus and the fissure recorded theta oscillation. The blue histogram shows the strength of the Tri-Synaptic Pathway's(TSP) influence on area CA1 over time, and the green histogram shows the Mono-Synaptic Pathway's(MSP) influence on CA1 on the same time line. The orange line represents the fissure recorded theta oscillation in reference to these histograms. Dotted lines show the points of maximum influence from either the TSP or MSP on CA1. B) Visual depiction of computational model shown at three sequential time points, with arrow weight highlighting the manipulated connection strengths at those time points; connections not depicted imply there was no modification of connection strength. The three time points of interest, theta peak, theta trough, and theta plus are shown with the influence of the MSP(shown in green) on CA1 strong at theta's trough, the influence of the TSP(shown in blue) on CA1 strong at the peak, and the influence of  on  (shown in black), as well as  to CA1 strong during theta plus. The transition from theta plus to the following theta trough is shown in the far right network.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3675133&req=5

pcbi-1003067-g002: Diagram of relation to physiological data and computational model.Figure adapted from [6]A) Relation of current-source/sink analysis within subfields of the hippocampus and the fissure recorded theta oscillation. The blue histogram shows the strength of the Tri-Synaptic Pathway's(TSP) influence on area CA1 over time, and the green histogram shows the Mono-Synaptic Pathway's(MSP) influence on CA1 on the same time line. The orange line represents the fissure recorded theta oscillation in reference to these histograms. Dotted lines show the points of maximum influence from either the TSP or MSP on CA1. B) Visual depiction of computational model shown at three sequential time points, with arrow weight highlighting the manipulated connection strengths at those time points; connections not depicted imply there was no modification of connection strength. The three time points of interest, theta peak, theta trough, and theta plus are shown with the influence of the MSP(shown in green) on CA1 strong at theta's trough, the influence of the TSP(shown in blue) on CA1 strong at the peak, and the influence of on (shown in black), as well as to CA1 strong during theta plus. The transition from theta plus to the following theta trough is shown in the far right network.

Mentions: Figure 2A shows an illustration of hippocampal subfield dynamics in relation to the fissure recorded theta oscillation shown in red. This cartoon, derived from current source density analysis [6], [15], shows the current sinks into area CA1 alternatively originating from either area CA3 in blue or EC layers II and III in green. At the trough of fissure recorded theta, EC sources into CA1 are at their peak and area CA3 is at its minimum. This implies that EC has a strong influence over synaptic potentials within area CA1 at this time. At the peak of fissure recorded theta, CA3 sources are at their peak and EC influence has diminished. This again suggests that CA3 input to area CA1 is now the dominant influence, and EC is less so as compared to the trough of the theta oscillation.


Theta coordinated error-driven learning in the hippocampus.

Ketz N, Morkonda SG, O'Reilly RC - PLoS Comput. Biol. (2013)

Diagram of relation to physiological data and computational model.Figure adapted from [6]A) Relation of current-source/sink analysis within subfields of the hippocampus and the fissure recorded theta oscillation. The blue histogram shows the strength of the Tri-Synaptic Pathway's(TSP) influence on area CA1 over time, and the green histogram shows the Mono-Synaptic Pathway's(MSP) influence on CA1 on the same time line. The orange line represents the fissure recorded theta oscillation in reference to these histograms. Dotted lines show the points of maximum influence from either the TSP or MSP on CA1. B) Visual depiction of computational model shown at three sequential time points, with arrow weight highlighting the manipulated connection strengths at those time points; connections not depicted imply there was no modification of connection strength. The three time points of interest, theta peak, theta trough, and theta plus are shown with the influence of the MSP(shown in green) on CA1 strong at theta's trough, the influence of the TSP(shown in blue) on CA1 strong at the peak, and the influence of  on  (shown in black), as well as  to CA1 strong during theta plus. The transition from theta plus to the following theta trough is shown in the far right network.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3675133&req=5

pcbi-1003067-g002: Diagram of relation to physiological data and computational model.Figure adapted from [6]A) Relation of current-source/sink analysis within subfields of the hippocampus and the fissure recorded theta oscillation. The blue histogram shows the strength of the Tri-Synaptic Pathway's(TSP) influence on area CA1 over time, and the green histogram shows the Mono-Synaptic Pathway's(MSP) influence on CA1 on the same time line. The orange line represents the fissure recorded theta oscillation in reference to these histograms. Dotted lines show the points of maximum influence from either the TSP or MSP on CA1. B) Visual depiction of computational model shown at three sequential time points, with arrow weight highlighting the manipulated connection strengths at those time points; connections not depicted imply there was no modification of connection strength. The three time points of interest, theta peak, theta trough, and theta plus are shown with the influence of the MSP(shown in green) on CA1 strong at theta's trough, the influence of the TSP(shown in blue) on CA1 strong at the peak, and the influence of on (shown in black), as well as to CA1 strong during theta plus. The transition from theta plus to the following theta trough is shown in the far right network.
Mentions: Figure 2A shows an illustration of hippocampal subfield dynamics in relation to the fissure recorded theta oscillation shown in red. This cartoon, derived from current source density analysis [6], [15], shows the current sinks into area CA1 alternatively originating from either area CA3 in blue or EC layers II and III in green. At the trough of fissure recorded theta, EC sources into CA1 are at their peak and area CA3 is at its minimum. This implies that EC has a strong influence over synaptic potentials within area CA1 at this time. At the peak of fissure recorded theta, CA3 sources are at their peak and EC influence has diminished. This again suggests that CA3 input to area CA1 is now the dominant influence, and EC is less so as compared to the trough of the theta oscillation.

Bottom Line: We show that the differential phase relationships of hippocampal subfields within the overall theta rhythm enable a powerful form of error-driven learning, which results in significantly greater capacity, as shown in computer simulations.In one phase of the theta cycle, the bidirectional connectivity between CA1 and entorhinal cortex can be trained in an error-driven fashion to learn to effectively encode the cortical inputs in a compact and sparse form over CA1.Taken together, these new learning dynamics enable a much more robust, high-capacity model of hippocampal learning than was available previously under the classical Hebbian model.

View Article: PubMed Central - PubMed

Affiliation: Department of Psychology, University of Colorado Boulder, Boulder, Colorado, United States of America.

ABSTRACT
The learning mechanism in the hippocampus has almost universally been assumed to be Hebbian in nature, where individual neurons in an engram join together with synaptic weight increases to support facilitated recall of memories later. However, it is also widely known that Hebbian learning mechanisms impose significant capacity constraints, and are generally less computationally powerful than learning mechanisms that take advantage of error signals. We show that the differential phase relationships of hippocampal subfields within the overall theta rhythm enable a powerful form of error-driven learning, which results in significantly greater capacity, as shown in computer simulations. In one phase of the theta cycle, the bidirectional connectivity between CA1 and entorhinal cortex can be trained in an error-driven fashion to learn to effectively encode the cortical inputs in a compact and sparse form over CA1. In a subsequent portion of the theta cycle, the system attempts to recall an existing memory, via the pathway from entorhinal cortex to CA3 and CA1. Finally the full theta cycle completes when a strong target encoding representation of the current input is imposed onto the CA1 via direct projections from entorhinal cortex. The difference between this target encoding and the attempted recall of the same representation on CA1 constitutes an error signal that can drive the learning of CA3 to CA1 synapses. This CA3 to CA1 pathway is critical for enabling full reinstatement of recalled hippocampal memories out in cortex. Taken together, these new learning dynamics enable a much more robust, high-capacity model of hippocampal learning than was available previously under the classical Hebbian model.

Show MeSH