Neural Correlates of Optimal Multisensory Decision Making under Time-Varying Reliabilities with an Invariant Linear Probabilistic Population Code. Hou et al

Published: 10 August 2019| Version 1 | DOI: 10.17632/b8ybw4shv3.1
Contributors:
Han Hou, Qihao Zheng, Yuchen Zhao, Alexandre Pouget, Yong Gu

Description

Each .mat file stores each cell's raw data (e.g., spike trains and monkey's behavioral events) as well as pre-processed data (e.g., smoothed PSTHs for different conditions). All experimental figures can be reproduced by running the Matlab GUI in the GitHub repository: /Labtools/HH_Tools/GROUP_GUI.fig.

Files

Steps to reproduce

Experimental figures (Figure 1-3): 1). Download all .mat file from Mendeley Data 2). Download Matlab code from /Labtools in the GitHub repo 3). Run GROUP_GUI.fig in Matlab and press the corresponding buttons to reproduce the figures. Modeling figures (Figure 4-7, S4-S6) 1). Download Matlab code from /Multisensory-PPC in the GitHub repo 2). Using /Multisensory-PPC/lip_HH.m to reproduce the network model for ilPPC (Figure 4-7, S5-S6) 3). Using /Multisensory-PPC/MST_info/InfoLossGroundTruthScan.m to reproduce Figure S4. Detailed instructions will be illustrated in the GitHub README session.

Institutions

Institute of Neuroscience Chinese Academy of Sciences, Universite de Geneve Departement des Neurosciences fondamentales, University of the Chinese Academy of Sciences

Categories

Neuroscience, Computational Neuroscience, Multimodal Interaction, Cognitive Neuroscience, Decision Making, Bayesian Inference

Licence