Learning shapes cortical dynamics to enhance integration of relevant sensory input

Chadwick, Angus; Khan, Adil G.; Poort, Jasper; Blot, Antonin; Hofer, Sonja B.; Mrsic-Flogel, Thomas D.; Sahani, Maneesh

doi:10.1016/j.neuron.2022.10.001

Cited by 15 publications

(14 citation statements)

References 77 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our findings favour the proposal that V1 integrates incoming input with intrinsic and top-down orbitofrontal and dorsal influences. They add to the body of evidence suggesting that V1 is involved in higher levels of visual function than is commonly assumed (e.g., Alink et al, 2010;Astorga et al, 2022;Ayzenshtat et al, 2012;Chadwick et al, 2022;Kazandjian et al, 2021;Koivisto et al, 2011;Kok & de Lange, 2014;Lamme, 1995;Lamme et al, 1998;Muckli, 2010;Noudoost & Moore, 2011;Rademaker et al, 2019;Roland et al, 2006;Super et al, 2001;Zipser et al, 1996). Because the retinotopic format of V1 coding is shared across processing levels, it offers a common platform for bottom-up and top-down communication (see Deco & Lee, 2004;Kravitz et al, 2013;Mumford, 1992).…”

Section: Discussionmentioning

confidence: 87%

Backward masking reveals coarse-to-fine dynamics in human V1

Schuurmans¹,

Bennett²,

Petras³

et al. 2023

Preprint

View full text Add to dashboard Cite

Natural images exhibit luminance variations aligned across a broad spectrum of spatial frequencies (SFs). It has been proposed that, at early stages of processing, the coarse signals carried by the low SF (LSF) of the visual input are sent rapidly from primary visual cortex (V1) to ventral, dorsal and frontal regions to form a coarse representation of the input, which is later sent back to V1 to guide the processing of fine-grained high SFs (i.e., HSF). We used functional resonance imaging (fMRI) to investigate the role of human V1 in the coarse-to-fine integration of visual input. We disrupted the processing of the coarse and fine content of full-spectrum human face stimuli via backward masking of selective SF ranges (LSFs: <1.75cpd and HSFs: >1.75cpd) at specific times (50, 83, 100 or 150ms). In line with coarse-to-fine proposals, we found that (1) the selective masking of stimulus LSF disrupted V1 activity in the earliest time window, and progressively decreased in influence, while (2) an opposite trend was observed for the masking of a stimulus' HSF. This pattern of activity was found in V1, as well as in ventral (i.e. the Fusiform Face area, FFA), dorsal and orbitofrontal regions. We additionally presented participants with contrast negated stimuli. While contrast negation significantly reduced response amplitudes in the FFA, as well as coupling between FFA and V1, coarse-to-fine dynamics were not affected by this manipulation. The fact that V1 response dynamics to strictly identical stimulus sets differed depending on the masked scale adds to growing evidence that V1 role goes beyond the early and quasi-passive transmission of visual information to the rest of the brain. It instead indicates that V1 may yield a 'spatially registered common forum' or 'blackboard' that integrates top-down inferences with incoming visual signals through its recurrent interaction with high-level regions located in the inferotemporal, dorsal and frontal regions.

show abstract

Section: Discussionmentioning

confidence: 87%

Backward masking reveals coarse-to-fine dynamics in human V1

Schuurmans¹,

Bennett²,

Petras³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…In symmetric networks, the most amplifying mode is aligned with the most persistent mode (Fig. 3b; top) 57,58 , and thus does not generate activity transients (Fig. 3c; top)—accounting for the simple pattern completion dynamics seen in classical attractor networks with symmetric connectivity 5,7,31,33–35,42–44 (Fig.…”

Section: Resultsmentioning

confidence: 99%

“…2g–l). Thus, there is a form of ‘speed–accuracy’ trade-off between whether inputs should use the most amplifying or persistent mode: if information is required immediately following stimulus offset, such as in a perceptual decision-making task 13,40,58 , inputs need to use the persistent mode. However, if there is a time delay until the information is needed, as is the case in all working memory tasks 2,60 , then the most amplifying mode becomes the optimal input direction.…”

Section: Resultsmentioning

confidence: 99%

“…However, we reasoned that neither of these standard cost functions may be appropriate for understanding PFC function. The cue-delay cost is well justified when stimuli need to be decoded potentially instantaneously after cue onset, and as such it is most relevant for sensory areas 58 . Conversely, the after-go-time cost may be most directly relevant for motor areas, by only requiring stable coding during the short response period 66 .…”

Section: Resultsmentioning

confidence: 99%

“…A number of recent studies of neural network dynamics have analysed the relationship between the direction of inputs and the magnitude of responses they evoke 52,56,61 . However, these studies have typically focused on networks with transient dynamics, such as those relevant for perception 58 , or motor control 52,61 . In particular, Ref.…”

Section: Discussionmentioning

confidence: 99%

See 2 more Smart Citations

Optimal information loading into working memory in prefrontal cortex explains dynamic coding

Stroud

Watanabe

Suzuki

et al. 2021

Preprint

View full text Add to dashboard Cite

Working memory involves the short-term maintenance of information and is critical in many tasks1. The neural circuit mechanisms underlying this information maintenance are thought to rely on persistent activities2,3 resulting from attractor dynamics4,5. However, how information is loaded into working memory for subsequent maintenance remains poorly understood. A pervasive assumption is that information loading requires inputs that are similar to the persistent activities expressed during maintenance6–9. Here, we show through mathematical analysis and numerical simulations that optimal inputs are instead largely orthogonal to persistent activities and naturally generate the rich transient dynamics characteristic of prefrontal cortex (PFC) during working memory10–17. By analysing recordings from monkeys performing a memory-guided saccade task, and using a novel, theoretically principled metric, we show that PFC exhibits the hallmarks of optimal information loading. Our theory unifies previous, seemingly conflicting theories of memory maintenance based on attractor8,9,18–20 or purely sequential dynamics21–23, and reveals a normative principle underlying the widely observed phenomenon of dynamic coding in PFC10–12,15–17,24. These results suggest that optimal information loading may be a key component of attractor dynamics characterising various cognitive functions and cortical areas, including long-term memory25,26 and navigation27,28 in the hippocampus, and decision making in the PFC19,20,29,30.

show abstract

Signatures of task learning in neural representations

Gurnani,

Cayco Gajic

2023

Current Opinion in Neurobiology

View full text Add to dashboard Cite

Learning shapes cortical dynamics to enhance integration of relevant sensory input

Cited by 15 publications

References 77 publications

Backward masking reveals coarse-to-fine dynamics in human V1

Backward masking reveals coarse-to-fine dynamics in human V1

Optimal information loading into working memory in prefrontal cortex explains dynamic coding

Signatures of task learning in neural representations

Contact Info

Product

Resources

About