Sensory processing and categorization in cortical and deep neural networks

Pinotsis, Dimitris A.; Siegel, Markus; Miller, Earl K.

doi:10.1016/j.neuroimage.2019.116118

Cited by 8 publications

(9 citation statements)

References 75 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…is called transfer function and predicts average firing rate, similar to activation functions in deep neural networks 59,60 . Also, X  is the rate-constant of postsynaptic filtering, δ is synaptic gain and η is the postsynaptic potential at which the half of the maximum firing rate is achieved, see e.g.…”

Section: Deep Neural Fields Describe Neural Ensemble Activitymentioning

confidence: 99%

“…In turn, correlation distances between DMs, known as deviations, express second order differences, that is, differences in pairwise differences in neural activity or electric fields in different brain areas for the same cued locations. We used deviations to test for significant correspondence between memory representations 60,87,88 .   and 300 degrees) hemifield form distinct clusters, shown by ellipses.…”

Section: The Same Memory Is Stored By Electric Fields In Different Br...mentioning

confidence: 99%

See 1 more Smart Citation

In vivo ephaptic coupling allows memory network formation

Pinotsis

Miller

2023

Preprint

Self Cite

View full text Add to dashboard Cite

It is increasingly clear that memories are distributed across multiple brain areas. Such “engram complexes” are important features of memory formation and consolidation. Here, we test the hypothesis that engram complexes are formed in part by bioelectric fields that sculpt and guide the neural activity and tie together the areas that participate in engram complexes. Like the conductor of an orchestra, the fields influence each musician or neuron and orchestrate the output, the symphony. Our results use the theory of synergetics, machine learning and data from a spatial delayed saccade task and provide evidence for in vivo ephaptic coupling in memory representations.

show abstract

Section: Deep Neural Fields Describe Neural Ensemble Activitymentioning

confidence: 99%

Section: The Same Memory Is Stored By Electric Fields In Different Br...mentioning

confidence: 99%

In vivo ephaptic coupling allows memory network formation

Pinotsis

Miller

2023

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…This paper follows upon our recent work that focused on groups of neurons that represent memories known as neural ensembles (Figure 1). In 60 we studied computations performed by neural ensembles during a flexible sensorimotor decision making task 61 . We showed that neural ensembles in the same brain area performed different computations based on the rule applied during each trial, although the stimulus processed was the same.…”

Section: Deep Neural Fields Describe Neural Ensemble Structure In a Holistic Fashionmentioning

confidence: 99%

“…This paper and 29 focus on the structure and biophysics of neural ensembles. In related work 60 , we also studied the computations performed by ensembles using deep neural networks and behavioural models.…”

Section: Deep Neural Fields Describe Neural Ensemble Structure In a Holistic Fashionmentioning

confidence: 99%

Beyond dimension reduction: Stable electric fields emerge from and allow representational drift

Pinotsis

Miller

2021

Preprint

Self Cite

View full text Add to dashboard Cite

It is known that the exact neurons maintaining a given memory (the neural ensemble) change from trial to trial. This raises the question of how the brain achieves stability in the face of this representational drift. Here, we demonstrate that this stability emerges at the level of the electric fields that arise from neural activity. The electric fields, in turn, can act as “guard rails” that funnel higher dimensional variable neural activity along stable lower dimensional routes. We show that electric fields carry information about working memory content. We obtained the latent space associated with each memory. We then confirmed the stability of the electric field by mapping the latent space to different cortical patches (that comprise a neural ensemble) and reconstructing information flow between patches. Stable electric fields can allow latent states to be transferred between brain areas, in accord with modern engram theory.

show abstract

“…Some theoretical models have been proposed to explain how the category is learned in the neural system ( 26 – 28 ). But most of models show categorical phenomena that are consistent with some behavioral results, without showing neural activity that encodes category information observed in the PFC or other brain areas ( 29 , 30 ).…”

Section: Introductionmentioning

confidence: 99%

Category learning in a recurrent neural network with reinforcement learning

Zhang

Pan²,

Wang³

2022

Front. Psychiatry

View full text Add to dashboard Cite

It is known that humans and animals can learn and utilize category information quickly and efficiently to adapt to changing environments, and several brain areas are involved in learning and encoding category information. However, it is unclear that how the brain system learns and forms categorical representations from the view of neural circuits. In order to investigate this issue from the network level, we combine a recurrent neural network with reinforcement learning to construct a deep reinforcement learning model to demonstrate how the category is learned and represented in the network. The model consists of a policy network and a value network. The policy network is responsible for updating the policy to choose actions, while the value network is responsible for evaluating the action to predict rewards. The agent learns dynamically through the information interaction between the policy network and the value network. This model was trained to learn six stimulus-stimulus associative chains in a sequential paired-association task that was learned by the monkey. The simulated results demonstrated that our model was able to learn the stimulus-stimulus associative chains, and successfully reproduced the similar behavior of the monkey performing the same task. Two types of neurons were found in this model: one type primarily encoded identity information about individual stimuli; the other type mainly encoded category information of associated stimuli in one chain. The two types of activity-patterns were also observed in the primate prefrontal cortex after the monkey learned the same task. Furthermore, the ability of these two types of neurons to encode stimulus or category information was enhanced during this model was learning the task. Our results suggest that the neurons in the recurrent neural network have the ability to form categorical representations through deep reinforcement learning during learning stimulus-stimulus associations. It might provide a new approach for understanding neuronal mechanisms underlying how the prefrontal cortex learns and encodes category information.

show abstract

Sensory processing and categorization in cortical and deep neural networks

Cited by 8 publications

References 75 publications

In vivo ephaptic coupling allows memory network formation

In vivo ephaptic coupling allows memory network formation

Beyond dimension reduction: Stable electric fields emerge from and allow representational drift

Category learning in a recurrent neural network with reinforcement learning

Contact Info

Product

Resources

About