Skip to Main content Skip to Navigation
Reports

Large visual neuron assemblies receptive fields estimation using a super-resolution approach

Abstract : One primary goal in analyzing sensory neurons’ recordings is to map the sensory space to the neural response, thus estimating the neuron’s receptive fields (RFs). For visual neurons, the classical method to estimate RFs is the Spike Triggered Average (STA). In short, STA consists estimate the average stimulus before each spike evoked by a white noise stimulus whose block size can be ad-hoc tuned to target one single neuron. However, this approach becomes impractical to deal with in large scale recordings of heterogeneous populations of neurons since no single block size can match all neurons. Here, we aim to overcome this limitation by leveraging super resolution techniques to extend STA’s scope. We defined a novel type of stimulus, the shifted white noise, by introducing random spatial shifts in the white noise stimulus. We evaluated this new stimulus thoroughly on both synthetic and real neuronal populations of size 216 and 4798, respectively. Considering the same target STA resolution, results across the population with synthetic case show that the average error using our stimulus was 1.7 times smaller than the error using the classical stimulus. We could map 2.3 times more neurons and cover a broader heterogeneity of RF sizes. For a single neuron, we show how it can be mapped after only one minute of stimulation, while after 11 minutes, this neuron was still not mapped with the classical one, which emphasizes the effectiveness of our method. Analogously, similar results were obtained with real neurons’ experiment. Considering the same target STA resolution, we mapped 18 times more RFs and we found a broader heterogeneity of RFs sizes (the kurtosis of the distribution of the RF sizes is 0.3 times smaller). Overall, the shifted white noise improves the RFs’ estimation in several ways. Our approach performs better at the single-cell level. RF estimation is independent of the neuron’s position relative to the stimulus and offers high-resolution. Our approach is stronger at the population-level. We get more RF with more neuronal variability. Our approach is faster, enabling experimentalists to get results in shorter stimulation time. Furthermore, this stimulus can also be used in other spike-triggered methods, extended to the time dimension, and adapted to other sensory modalities. Due to its design simplicity and strong results, we expect that soon the shifted white noise is used as a rule and allows revealing novelties in sensory analysis.
Document type :
Reports
Complete list of metadatas

https://hal.inria.fr/hal-03087009
Contributor : Pierre Kornprobst <>
Submitted on : Wednesday, December 23, 2020 - 10:59:51 AM
Last modification on : Monday, January 4, 2021 - 8:59:57 AM

File

RR-9383.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-03087009, version 1

Citation

Daniela Pamplona, Gerrit Hilgen, Matthias Hennig, Bruno Cessac, Evelyne Sernagor, et al.. Large visual neuron assemblies receptive fields estimation using a super-resolution approach. [Research Report] RR-9383, Inria - Sophia antipolis. 2020. ⟨hal-03087009⟩

Share

Metrics

Record views

74

Files downloads

285