Limits...
Choice-correlated activity fluctuations underlie learning of neuronal category representation.

Engel TA, Chaisangmongkon W, Freedman DJ, Wang XJ - Nat Commun (2015)

Bottom Line: To elucidate the neural mechanisms underlying categorization, we constructed a cortical circuit model that is capable of learning a motion categorization task through reward-dependent plasticity.In the model, choice probability and task-specific interneuronal correlations emerge from plasticity of top-down projections from decision neurons.Beyond demonstrating a circuit mechanism for categorization, the present work suggests a key role of plastic top-down feedback in simultaneously shaping both neural tuning and correlated neural variability.

View Article: PubMed Central - PubMed

Affiliation: 1] Department of Neurobiology, Yale University School of Medicine, Kavli Institute for Neuroscience, 333 Cedar Street, New Haven, Connecticut 06510, USA [2] Department of Bioengineering, Stanford University, 318 Campus Drive, Stanford, California 94305, USA.

ABSTRACT
The ability to categorize stimuli into discrete behaviourally relevant groups is an essential cognitive function. To elucidate the neural mechanisms underlying categorization, we constructed a cortical circuit model that is capable of learning a motion categorization task through reward-dependent plasticity. Here we show that stable category representations develop in neurons intermediate to sensory and decision layers if they exhibit choice-correlated activity fluctuations (choice probability). In the model, choice probability and task-specific interneuronal correlations emerge from plasticity of top-down projections from decision neurons. Specific model predictions are confirmed by analysis of single-neuron activity from the monkey parietal cortex, which reveals a mixture of directional and categorical tuning, and a positive correlation between category selectivity and choice probability. Beyond demonstrating a circuit mechanism for categorization, the present work suggests a key role of plastic top-down feedback in simultaneously shaping both neural tuning and correlated neural variability.

No MeSH data available.


Related in: MedlinePlus

Association neurons in the network with feedback, but not in the network without feedback, exhibit choice-correlated fluctuations.(a) In the network without feedback, CP is close to 0.5 in all association neurons and does not change throughout learning. (b) In the network with feedback, CP is randomly scattered around 0.5 before learning (grey dots), but a bimodal profile of CP develops after a short period of learning (500 trials, green dots), such that CP>0.5 in neurons with preferred directions in category C1, and CP<0.5 in neurons with preferred directions in category C2. CP (y axis) is plotted for all association neurons labeled by their initial preferred direction (x axis) before (grey dots) and after (olive dots) learning. Histograms to the right show the corresponding CP distributions.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4382677&req=5

f6: Association neurons in the network with feedback, but not in the network without feedback, exhibit choice-correlated fluctuations.(a) In the network without feedback, CP is close to 0.5 in all association neurons and does not change throughout learning. (b) In the network with feedback, CP is randomly scattered around 0.5 before learning (grey dots), but a bimodal profile of CP develops after a short period of learning (500 trials, green dots), such that CP>0.5 in neurons with preferred directions in category C1, and CP<0.5 in neurons with preferred directions in category C2. CP (y axis) is plotted for all association neurons labeled by their initial preferred direction (x axis) before (grey dots) and after (olive dots) learning. Histograms to the right show the corresponding CP distributions.

Mentions: This general principle explains both the fast associative learning and slower behavioural improvements in our model. Since activities of decision neurons directly represent the model’s choices, the magnitude of their CP is large; hence, the synapses of decision neurons change rapidly towards increasing expected reward, underpinning fast associative learning. In the network with feedback, CP arises via feedback from the decision circuit, which produces multiplicative rate modulations in association neurons3031 (Supplementary Fig. 2). Initially, CP is scattered around 0.5; however, when feedback connections become structured (~500 trials), neurons receiving stronger input from the C1 (C2) decision population fire at higher rates when C1 (C2) choices are made and exhibit CP>0.5 (CP<0.5, Fig. 6b). The magnitude of CP is smaller in association than in decision neurons; therefore, the tuning changes of association neurons and ensuing behavioural improvements happen more slowly than associative learning. In the network without feedback, CP≈0.5 in all association neurons and at all learning stages (Fig. 6a), because local noise in the decision circuit—required to attain realistic behavioural performance in the categorization task—diminishes the influence of association neurons' rate fluctuations on choices (see Supplementary Note 4 for details). Resulting unstructured synaptic changes lead to deterioration of tuning and behavioural performance. Regardless of which mechanism—feedforward or feedback—is more plausible for generating CP in real neurons, our results demonstrate the significance of CP for reward-dependent learning.


Choice-correlated activity fluctuations underlie learning of neuronal category representation.

Engel TA, Chaisangmongkon W, Freedman DJ, Wang XJ - Nat Commun (2015)

Association neurons in the network with feedback, but not in the network without feedback, exhibit choice-correlated fluctuations.(a) In the network without feedback, CP is close to 0.5 in all association neurons and does not change throughout learning. (b) In the network with feedback, CP is randomly scattered around 0.5 before learning (grey dots), but a bimodal profile of CP develops after a short period of learning (500 trials, green dots), such that CP>0.5 in neurons with preferred directions in category C1, and CP<0.5 in neurons with preferred directions in category C2. CP (y axis) is plotted for all association neurons labeled by their initial preferred direction (x axis) before (grey dots) and after (olive dots) learning. Histograms to the right show the corresponding CP distributions.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4382677&req=5

f6: Association neurons in the network with feedback, but not in the network without feedback, exhibit choice-correlated fluctuations.(a) In the network without feedback, CP is close to 0.5 in all association neurons and does not change throughout learning. (b) In the network with feedback, CP is randomly scattered around 0.5 before learning (grey dots), but a bimodal profile of CP develops after a short period of learning (500 trials, green dots), such that CP>0.5 in neurons with preferred directions in category C1, and CP<0.5 in neurons with preferred directions in category C2. CP (y axis) is plotted for all association neurons labeled by their initial preferred direction (x axis) before (grey dots) and after (olive dots) learning. Histograms to the right show the corresponding CP distributions.
Mentions: This general principle explains both the fast associative learning and slower behavioural improvements in our model. Since activities of decision neurons directly represent the model’s choices, the magnitude of their CP is large; hence, the synapses of decision neurons change rapidly towards increasing expected reward, underpinning fast associative learning. In the network with feedback, CP arises via feedback from the decision circuit, which produces multiplicative rate modulations in association neurons3031 (Supplementary Fig. 2). Initially, CP is scattered around 0.5; however, when feedback connections become structured (~500 trials), neurons receiving stronger input from the C1 (C2) decision population fire at higher rates when C1 (C2) choices are made and exhibit CP>0.5 (CP<0.5, Fig. 6b). The magnitude of CP is smaller in association than in decision neurons; therefore, the tuning changes of association neurons and ensuing behavioural improvements happen more slowly than associative learning. In the network without feedback, CP≈0.5 in all association neurons and at all learning stages (Fig. 6a), because local noise in the decision circuit—required to attain realistic behavioural performance in the categorization task—diminishes the influence of association neurons' rate fluctuations on choices (see Supplementary Note 4 for details). Resulting unstructured synaptic changes lead to deterioration of tuning and behavioural performance. Regardless of which mechanism—feedforward or feedback—is more plausible for generating CP in real neurons, our results demonstrate the significance of CP for reward-dependent learning.

Bottom Line: To elucidate the neural mechanisms underlying categorization, we constructed a cortical circuit model that is capable of learning a motion categorization task through reward-dependent plasticity.In the model, choice probability and task-specific interneuronal correlations emerge from plasticity of top-down projections from decision neurons.Beyond demonstrating a circuit mechanism for categorization, the present work suggests a key role of plastic top-down feedback in simultaneously shaping both neural tuning and correlated neural variability.

View Article: PubMed Central - PubMed

Affiliation: 1] Department of Neurobiology, Yale University School of Medicine, Kavli Institute for Neuroscience, 333 Cedar Street, New Haven, Connecticut 06510, USA [2] Department of Bioengineering, Stanford University, 318 Campus Drive, Stanford, California 94305, USA.

ABSTRACT
The ability to categorize stimuli into discrete behaviourally relevant groups is an essential cognitive function. To elucidate the neural mechanisms underlying categorization, we constructed a cortical circuit model that is capable of learning a motion categorization task through reward-dependent plasticity. Here we show that stable category representations develop in neurons intermediate to sensory and decision layers if they exhibit choice-correlated activity fluctuations (choice probability). In the model, choice probability and task-specific interneuronal correlations emerge from plasticity of top-down projections from decision neurons. Specific model predictions are confirmed by analysis of single-neuron activity from the monkey parietal cortex, which reveals a mixture of directional and categorical tuning, and a positive correlation between category selectivity and choice probability. Beyond demonstrating a circuit mechanism for categorization, the present work suggests a key role of plastic top-down feedback in simultaneously shaping both neural tuning and correlated neural variability.

No MeSH data available.


Related in: MedlinePlus