New perspectives on dimensionality and variability from large-scale cortical dynamics

Engel, Tatiana A.; Steinmetz, Nicholas A.

doi:10.1016/j.conb.2019.09.003

Cited by 44 publications

(35 citation statements)

References 96 publications

(120 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Cortical networks encode information in a multidimensional coordinate system such that, contingent upon the route of activation, that is, behavior, activation weights in one dimension have more gain relative to another dimension but remain coregistered through homomorphism with one another. 6 Through neural gain modulation, representations that are more relevant (i.e., have higher likelihoods) in a given context can dominate and guide behavior, giving the system the flexibility needed to dynamically amplify aspects of representations in relation to the sensory context and behavioral goal (see Engel & Steinmetz, 2019, for a review). In such a coding scheme for language, knowledge of the lexicon, and grammatical, semantic, and contextual knowledge can be "shared" across modalities and recruited during the assembly of representations for articulation, as well as during the perceptual inference and generation of structures during comprehension.…”

Section: Concrete Examples Of Gain Modulation For Coordinate Transformentioning

confidence: 99%

A Compositional Neural Architecture for Language

Martin

2020

Journal of Cognitive Neuroscience

175

View full text Add to dashboard Cite

Hierarchical structure and compositionality imbue human language with unparalleled expressive power and set it apart from other perception–action systems. However, neither formal nor neurobiological models account for how these defining computational properties might arise in a physiological system. I attempt to reconcile hierarchy and compositionality with principles from cell assembly computation in neuroscience; the result is an emerging theory of how the brain could convert distributed perceptual representations into hierarchical structures across multiple timescales while representing interpretable incremental stages of (de)compositional meaning. The model's architecture—a multidimensional coordinate system based on neurophysiological models of sensory processing—proposes that a manifold of neural trajectories encodes sensory, motor, and abstract linguistic states. Gain modulation, including inhibition, tunes the path in the manifold in accordance with behavior and is how latent structure is inferred. As a consequence, predictive information about upcoming sensory input during production and comprehension is available without a separate operation. The proposed processing mechanism is synthesized from current models of neural entrainment to speech, concepts from systems neuroscience and category theory, and a symbolic-connectionist computational model that uses time and rhythm to structure information. I build on evidence from cognitive neuroscience and computational modeling that suggests a formal and mechanistic alignment between structure building and neural oscillations, and moves toward unifying basic insights from linguistics and psycholinguistics with the currency of neural computation.

show abstract

Section: Concrete Examples Of Gain Modulation For Coordinate Transformentioning

confidence: 99%

A Compositional Neural Architecture for Language

Martin

2020

Journal of Cognitive Neuroscience

175

View full text Add to dashboard Cite

show abstract

“…Recent advances in electrophysiology and optical imaging techniques enable to study in awake behaving animals the persistence over time of neuronal coding properties, such as the tuning of neurons to specific stimuli (Rokni et al ., 2007; Tolias et al ., 2007; Bondar et al ., 2009; Andermann, Kerlin and Reid, 2010; Huber et al ., 2012; Ziv et al ., 2013; Peron et al ., 2015; Poort et al ., 2015; Okun et al ., 2016; Dhawale et al ., 2017; Jun et al ., 2017). Some of these studies exposed a substantial degree of variability in neuronal responses to the same stimuli over timescales spanning minutes to weeks, prompting neuroscientists to question the naïve assumption that stable neuronal codes are essential for stable brain functionality (Tolhurst, Movshon and Dean, 1983; Arieli et al ., 1996; Rokni et al ., 2007; Faisal, Selen and Wolpert, 2008; Minerbi et al ., 2009; Cohen and Maunsell, 2010; Huber et al ., 2012; Ziv et al ., 2013; Lütcke, Margolis and Helmchen, 2013; Montijn, Goltstein and Pennartz, 2015; Rubin et al ., 2015; Schölvinck et al ., 2015; Rose et al ., 2016; Chambers and Rumpel, 2017; Clopath et al ., 2017; Dhawale et al ., 2017; Driscoll et al ., 2017; Engel and Steinmetz, 2019; Rule et al ., 2020; Sheintuch et al ., 2020).…”

Section: Introductionmentioning

confidence: 99%

Representational drift in the mouse visual cortex

Deitch

Rubin

Ziv

2020

Preprint

View full text Add to dashboard Cite

Neuronal representations in the hippocampus and related structures gradually change over time despite no changes in the environment or behavior. The extent to which such ‘representational drift’ occurs in sensory cortical areas and whether the hierarchy of information flow across areas affects neural-code stability have remained elusive. Here, we address these questions by analyzing large-scale optical and electrophysiological recordings from six visual cortical areas in behaving mice that were repeatedly presented with the same natural movies. We found representational drift over timescales spanning minutes to days across multiple visual areas. The drift was driven mostly by changes in individual cells’ activity rates, while their tuning changed to a lesser extent. Despite these changes, the structure of relationships between the population activity patterns remained stable and stereotypic, allowing robust maintenance of information over time. Such population-level organization may underlie stable visual perception in the face of continuous changes in neuronal responses.

show abstract

“…The second regime is where the firing rates are no longer dynamically stable, and large rate fluctuations are produced through internal interactions (as we will formalize below). In past work 47 we showed how model networks in this second regime produced the low-dimensional, population-wide shared variability characteristic of a variety of cortical areas 9–11,47,66–69 . In this section we consider how networks in this second regime transfer information across layers.…”

Section: Resultsmentioning

confidence: 91%

“…A prominent feature of cortical response to sensory stimuli is that neuronal activity varies significantly across presentation trials 1,2 , even when efforts are taken to control or account for variable animal behavior 3–6 . A component of this variability is coordinated among neurons in a brain area, often leading to shared fluctuations in spiking activity 7–11 . How stimulus processing is affected by this large, population-wide neuronal variability is a longstanding question in both experimental and theoretical neuroscience communities.…”

Section: Introductionmentioning

confidence: 99%

Internally generated population activity in cortical networks hinders information transmission

Huang

Pouget

Doiron

2020

Preprint

View full text Add to dashboard Cite

4How neuronal variability impacts neuronal codes is a central question in systems neuroscience, often 5 with complex and model dependent answers. Most population models are parametric, with a tacitly 6 assumed structure of neuronal tuning and population-wide variability. While these models provide 7 key insights, they purposely divorce any mechanistic relationship between trial average and trial vari-8 able neuronal activity. By contrast, circuit based models produce activity with response statistics that 9 are reflection of the underlying circuit structure, and thus any relations between trial averaged and 10 trial variable activity are emergent rather than assumed. In this work, we study information transfer 11 in networks of spatially ordered spiking neuron models with strong excitatory and inhibitory interac-12 tions, capable of producing rich population-wide neuronal variability. Motivated by work in the visual 13 system we embed a columnar stimulus orientation map in the network and measure the population 14 estimation of an orientated input. We show that the spatial structure of feedforward and recurrent 15 connectivity are critical determinants for population code performance. In particular, when network 16 wiring supports stable firing rate activity then with a sufficiently large number of decoded neurons all 17 1 available stimulus information is transmitted. However, if the inhibitory projections place network ac-18 tivity in a pattern forming regime then the population-wide dynamics compromise information flow. 19 In total, network connectivity determines both the stimulus tuning as well as internally generated 20 population-wide fluctuations and thereby dictates population code performance in complicated ways 21 where modeling efforts provide essential understanding. 23A prominent feature of cortical response to sensory stimuli is that neuronal activity varies significantly 24 across presentation trials 1,2 , even when efforts are taken to control or account for variable animal 25 behavior 3-6 . A component of this variability is coordinated among neurons in a brain area, often 26 leading to shared fluctuations in spiking activity [7][8][9][10][11] . How stimulus processing is affected by this large, 27 population-wide neuronal variability is a longstanding question in both experimental and theoretical 28 neuroscience communities. 29 Recording from neuronal populations while simultaneously monitoring an animal's behavior dur-30 ing a structured task offers a glimpse into how neuronal activity supports computation. Correlations 31 between spike counts from pairs of neurons in response to repeated stimulus, often referred to as 32 noise correlations, are modulated by a variety of cognitive factors that are known to affect task perfor-33 mance 4 . For example, noise correlations decrease with animal arousal 12,13 or task engagement 14 . In 34 the visual pathway noise correlations are decreased when spatial attention is directed into the recep-35 tive field of a recorded population 15,16 ...

show abstract

New perspectives on dimensionality and variability from large-scale cortical dynamics

Cited by 44 publications

References 96 publications

A Compositional Neural Architecture for Language

A Compositional Neural Architecture for Language

Representational drift in the mouse visual cortex

Internally generated population activity in cortical networks hinders information transmission

Contact Info

Product

Resources

About