# Explaining the Human Visual Brain Challenge 2019 -- receptive fields and surrogate features

## TL;DR

This paper reviews submissions to a challenge focused on aligning neural network features with human brain data from fMRI and MEG, emphasizing receptive fields, surrogate features, and RDM construction.

## Contribution

It introduces methods for optimizing receptive field granularity, creating surrogate features with Multidimensional Scaling, and discusses RDM construction nuances.

## Key findings

- Optimal receptive field granularity improves RDM similarity.
- Surrogate features via Multidimensional Scaling enhance model-brain alignment.
- Attention to RDM construction details is crucial for accurate comparisons.

## Abstract

In this paper I review the submission to the Explaining the Human Visual Brain Challenge 2019 in both the fMRI and MEG tracks. The goal was to construct neural network features which generate the so-called representational dissimilarity matrix (RDM) which is most similar to the one extracted from fMRI and MEG data upon viewing a set of images. I review exploring the optimal granularity of the receptive field, a construction of intermediate surrogate features using Multidimensional Scaling and modelling them using neural network features. I also point out some peculiarities of the RDM construction which have to be taken into account.