Skip to main content
Dryad

Anonymized source data files for figures in: Recurrent processes support a cascade of hierarchical decisions

Cite this dataset

Gwilliams, Laura; King, Jean-Remi (2020). Anonymized source data files for figures in: Recurrent processes support a cascade of hierarchical decisions [Dataset]. Dryad. https://doi.org/10.5061/dryad.70rxwdbtr

Abstract

Perception depends on a complex interplay between feedforward and recurrent processing. Yet, while the former has been extensively characterized, the computational organization of the latter remains largely unknown. Here, we use magneto-encephalography to localize, track and decode the feedforward and recurrent processes of reading, as elicited by letters and digits whose level of ambiguity was parametrically manipulated. We first confirm that a feedforward response propagates through the ventral and dorsal pathways within the first 200 ms. The subsequent activity is distributed across temporal, parietal and prefrontal cortices, which sequentially generate five levels of representations culminating in action-specific motor signals. Our decoding analyses reveal that both the content and the timing of these brain responses are best explained by a hierarchy of recurrent neural assemblies, which both maintain and broadcast increasingly rich representations. Together, these results show how recurrent processes generate, over extended time periods, a cascade of decisions that ultimately accounts for subjects' perceptual reports and reaction times.

Methods

The data were collected at NeuroSpin, Paris, France, with a CTF magneto-encraphalogram. Please see the manuscript for details on the experimental task.

Fig1: The grand average neural response to the visual stimuli, averaged over trials and subjects.

Fig2: Decoding performance across time (.npy files) and space (.stc files).

Fig4: Temporal generalisation decoding performance for each feature.

Fig 5: Re-aligned temporal generalisation matriced for each feature, and corresponding by-trial reaction times.

Fig 6: Probabalistic output of the classifier over time, for the perceived letter/digit category (perceptual decision) and the motor action (motor decision).

Usage notes

The files are of two formats: .npy and .stc.

 

.npy files can be read using the Numpy module in python, e.g.:

import numpy as np

data = np.load('file_name.npy')

https://numpy.org/doc/stable/reference/generated/numpy.load.html

 

 

.stc files can be read using the MNE module in python, e.g.:

from mne import read_source_estimate

stc = read_source_estimate('stc_name-lh.stc')

note that reading in the data from just one hemisphere file will automatically read the data for the other one too.

https://mne.tools/stable/generated/mne.read_source_estimate.html

Funding

William Orr Dingwall Foundation, Award: Dissertation Fellowship

European Commission, Award: 660086

Fondation Bettencourt Schueller, Award: Bettencourt-Schueller Foundation

Fondation Roger de Spoelberch, Award: Fondation Roger de Spoelberch

Philippe Foundation, Award: Philippe Foundation

National Cancer Institute, Award: R01DC05660

Abu Dhabi Institute Grant, Award: G1001