In EDS ansehen

Competition and cooperation of assembly sequences in recurrent neural networks.

Gespeichert in:
Bibliographische Detailangaben
Titel: Competition and cooperation of assembly sequences in recurrent neural networks.
Autoren: Stöber, Tristan Manfred1,2,3 (AUTHOR) tristan.stoeber@posteo.net, Lehr, Andrew Benjamin2,4 (AUTHOR), Nikzad, Arash5 (AUTHOR), Ganjtabesh, Mohammad5 (AUTHOR), Fyhn, Marianne2,6 (AUTHOR), Kumar, Arvind7,8 (AUTHOR)
Quelle: PLoS Computational Biology. 9/12/2025, Vol. 21 Issue 9, p1-22. 22p.
Schlagwörter: *NEURAL circuitry, *COOPERATIVENESS, *NEURAL codes, *ARTIFICIAL neural networks
Abstract: Neural activity sequences are ubiquitous in the brain and play pivotal roles in functions such as long-term memory formation and motor control. While conditions for storing and reactivating individual sequences have been thoroughly characterized, it remains unclear how multiple sequences may interact when activated simultaneously in recurrent neural networks. This question is especially relevant for weak sequences, comprised of fewer neurons, competing against strong sequences. Using a non-linear rate -based and a spiking model with discrete, pre-configured assemblies, we demonstrate that weak sequences can compensate for their competitive disadvantage either by increasing excitatory connections between subsequent assemblies or by cooperating with other co-active sequences. Further, our models suggest that such cooperation can negatively affect sequence speed unless subsequently active assemblies are paired. Our analysis characterizes the conditions for successful sequence progression in isolated, competing, and cooperating assembly sequences, and identifies the distinct contributions of recurrent and feed-forward projections. This proof-of-principle study shows how even disadvantaged sequences can be prioritized for reactivation, a process which has recently been implicated in hippocampal memory processing. Author summary: While competition and cooperation are well studied phenomena in various domains of life, the conditions under which they occur in neuronal circuits require further investigation. A central open question is what allows a neural activity sequence to reactivate when competing sequences are present. To address this knowledge gap, we built two types of simple network models: rate-based models and more biological-plausible spiking models. Each model contained up to three sequences, composed of discrete neuronal assemblies connected by feed-forward and recurrent projections. We found that a) both feed-forward and recurrent projections play crucial roles for a sequence to withstand competition, b) weak sequences can overcome their competitive disadvantage through cooperation, c) Hebbian cooperation can however negatively affect sequence speed, d) unless cooperative projections are shifted forward in time. Despite their simplicity, our rate-based and spiking models allow us to make predictions on network properties and plasticity mechanisms in brain regions involved in sequence processing. Given the ubiquitous role of neural activity sequences for information processing, a deeper comprehension of their dynamics will help to unravel basic and pathological brain functions. [ABSTRACT FROM AUTHOR]
Datenbank: Academic Search Index
Beschreibung
Abstract:Neural activity sequences are ubiquitous in the brain and play pivotal roles in functions such as long-term memory formation and motor control. While conditions for storing and reactivating individual sequences have been thoroughly characterized, it remains unclear how multiple sequences may interact when activated simultaneously in recurrent neural networks. This question is especially relevant for weak sequences, comprised of fewer neurons, competing against strong sequences. Using a non-linear rate -based and a spiking model with discrete, pre-configured assemblies, we demonstrate that weak sequences can compensate for their competitive disadvantage either by increasing excitatory connections between subsequent assemblies or by cooperating with other co-active sequences. Further, our models suggest that such cooperation can negatively affect sequence speed unless subsequently active assemblies are paired. Our analysis characterizes the conditions for successful sequence progression in isolated, competing, and cooperating assembly sequences, and identifies the distinct contributions of recurrent and feed-forward projections. This proof-of-principle study shows how even disadvantaged sequences can be prioritized for reactivation, a process which has recently been implicated in hippocampal memory processing. Author summary: While competition and cooperation are well studied phenomena in various domains of life, the conditions under which they occur in neuronal circuits require further investigation. A central open question is what allows a neural activity sequence to reactivate when competing sequences are present. To address this knowledge gap, we built two types of simple network models: rate-based models and more biological-plausible spiking models. Each model contained up to three sequences, composed of discrete neuronal assemblies connected by feed-forward and recurrent projections. We found that a) both feed-forward and recurrent projections play crucial roles for a sequence to withstand competition, b) weak sequences can overcome their competitive disadvantage through cooperation, c) Hebbian cooperation can however negatively affect sequence speed, d) unless cooperative projections are shifted forward in time. Despite their simplicity, our rate-based and spiking models allow us to make predictions on network properties and plasticity mechanisms in brain regions involved in sequence processing. Given the ubiquitous role of neural activity sequences for information processing, a deeper comprehension of their dynamics will help to unravel basic and pathological brain functions. [ABSTRACT FROM AUTHOR]
ISSN:1553734X
DOI:10.1371/journal.pcbi.1013403