2022
DOI: 10.1101/2022.04.11.487906
|View full text |Cite
Preprint
|
Sign up to set email alerts
|

Between-area communication through the lens of within-area neuronal dynamics

Abstract: Cognition arises through the interaction of distributed, but connected, brain regions. Each region can exhibit neuronal dynamics ranging from asynchronous spiking to richly patterned spatio-temporal activity, where coordinated trial-to-trial fluctuations within the population can be described by a small number of shared latent variables. Even though recent technological advances have allowed simultaneous recordings from multiple brain areas, it is still unknown how diverse and complex within-area neuronal dyna… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
5

Citation Types

2
4
0

Year Published

2022
2022
2024
2024

Publication Types

Select...
4
1

Relationship

0
5

Authors

Journals

citations
Cited by 5 publications
(6 citation statements)
references
References 89 publications
(167 reference statements)
2
4
0
Order By: Relevance
“…Our findings comport with recent theoretical results suggesting that local circuit dynamics can be the primary determinant of firing patterns in the presence of substantial across-region input ( 40, 41 ). Across-region connections may instead function to coordinate firing patterns across regions or modulate higher-order firing pattern features, like their autocorrelation ( 55 ).…”
Section: Discussionsupporting
confidence: 91%
See 1 more Smart Citation
“…Our findings comport with recent theoretical results suggesting that local circuit dynamics can be the primary determinant of firing patterns in the presence of substantial across-region input ( 40, 41 ). Across-region connections may instead function to coordinate firing patterns across regions or modulate higher-order firing pattern features, like their autocorrelation ( 55 ).…”
Section: Discussionsupporting
confidence: 91%
“…The propagation of local activity in the service of this systemic pattern generation could be another constraint, as a number of recent observations suggest that local circuit dynamics govern motor cortical firing ( 56, 57 ) to some extent ( 58 ). Our results underscore the recognized need ( 40 ) for developing theory about across-region interactions.…”
Section: Discussionsupporting
confidence: 68%
“…Our final contribution is to incorporate the empirical findings within A1 and PFC in a multi-area network that postulates their interactions through low-rank communication subspaces. Previous work modeling communication subspaces have focused on noise correlations in spontaneous activity and feedforward interactions (Gozel and Doiron 2022; Thivierge and Pilzak 2022); see also (Perich, Gallego, and L. E. Miller 2018) for a model of ‘output-null’ subspaces in the context of motor preparation. In contrast, we now propose a multi-region network with interactions in both feedforward and feedback directions, where PFC acts as a controller of A1, dynamically selecting the appropriate communication subspace for the ongoing context.…”
Section: Discussionmentioning
confidence: 99%
“…In this work, we opted to model the simplest hypothesis that could explain our findings within A1 and PFC. Regardless, this model is to our knowledge the first neural implementation of the communication subspace hypothesis (Semedo, Zandvakili, et al 2019) that performs a cognitive task (but see Gozel and Doiron 2022; Thivierge and Pilzak 2022). While future theoretical work will be necessary to fully flesh out the implications of the communication subspace hypothesis (Semedo, Jasper, et al 2022; Semedo, Zandvakili, et al 2019), our model reveals several interesting insights.…”
Section: Discussionmentioning
confidence: 99%
“…Indeed, the predominant explanation for anatomical organization of the brain has focused on minimizing wiring costs while maximizing adaptive topological features (Bullmore & Sporns, 2012; Zhou et al, 2022). As such, we have seen considerations of space implemented in functional feedforward neural network models (Gozel & Doiron, 2022; Huang et al, 2019; Lee et al, 2020). Here we instantiated our core hypothesis mathematically within the seRNN model by providing two challenges to RNNs during supervised learning: (1) long connections should be minimized where possible – reflective of their metabolic cost (Kaiser & Hilgetag, 2006; Sporns, 2011), and (2) connections can only change their weights as a function of their underlying communication – reflective of signal propagation between neuronal units (Betzel et al, 2022; Seguin, Jedynak, et al, 2022; Seguin, Mansour L, et al, 2022; Shimono & Hatano, 2018).…”
Section: Discussionmentioning
confidence: 99%