---
layout: default
title: "methods"
comments: true
---



# <span style="color:lightblue">Methods</span>

<hr style="border: 1px solid black; width:100%;"></hr>

The methods presented here are a work in progress. For implementation inquiries or clarifications, please do not hesitate to contact me.

## Download the HCP 7T-fMRI retinotopy dataset

* Fetch the HCP data using [AWS](https://wiki.humanconnectome.org/display/PublicData/How+To+Connect+to+Connectome+Data+via+AWS) and Python ([Jupyter notebook](https://github.com/nicogravel/retNet_EC/blob/main/notebooks/HCP_7T_fetch.ipynb))


## Grouping V1, V2, and V3 data into foveal and para-foveal ROIs

* Define the ROIs using the average pRF parameters and the Wang-Kastner ([Jupyter notebook](https://github.com/nicogravel/retNet_EC/blob/main/notebooks/HCP_ECVP_quadrants.ipynb))

## Modelling the propagation of BOLD activity across V1, V2 and V3 using a multivariate Ornstein-Uhlenbeck network model

* Fit **network model** ([Jupyter notebook](https://github.com/nicogravel/retNet_EC/blob/main/notebooks/HCP_7T_ret_quad_NDNetwork_fit.ipynb))

* Statistical assessment of the model's effective connectivity (EC) and its task-dependent changes ([Jupyter notebook](https://github.com/nicogravel/retNet_EC/blob/main/notebooks/HCP_7T_ret_quad_NDNetwork_stat.ipynb))

* Visualize the structure of EC interactions across ROIs using a [2D representation of the foveal confluence](https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1000651) ([Matlab script](https://github.com/nicogravel/retNet_EC/blob/main/scripts/plot_EC.m))

* Group the model's parameters into larger ROIs for interpretation ([Matlab script](https://github.com/nicogravel/retNet_EC/blob/main/scripts/plot_sigma.m))

<br>
<br>
<br>
<br>
<br>


```{disqus}
```