Analyses of the individual electric field magnitudes with Roast.

Overview

Aloi Davide - PhD Student (UoB)

Analysis of electric field magnitudes (wp2a dataset only at the moment) and correlation analysis with Dynamic Causal Modelling (DCM) results.

The goal of these analyses is to establish whether there is a relationship between single-subject electric field (E-field) magnitudes generated with the ROAST pipeline (Huang et al., 2019) and changes in effective connectivity within the motor network, derived using DCM and parametric empirical bayes (PEB).

The two analyses are:

  1. Correlation analysis between E-field magnitude - medians and max values - (or the current density?) in the motor cortex (M1) and Thalamus (Th) with self- / between-connectivities (M1 and Th only?) as derived from the DCM. e.g. Indahlastari et al. (2021). At the moment I am correlating e-field measures only with DCM measures derived from the contrast pre vs post Day-1 anodal only. However, I should also correlate those e-field measures with DCM measures derived from the contrast pre vs post Day-1 sham. I expect to find correlations between e-field measures and DCM measures for the anodal condition but not for sham.
  2. Pattern-recognition analysis using support vector machine (SVM) learning algorithm on MRI-derived tDCS current models to provide classification of tDCS treatment response (as reflected by increased M1-TH or TH-M1 connectivity or whatever other measure we decide). e.g. Albizu et al. (2020). The question here is: can we classify people who had an increase in thalamo-cortical connectivity using features from the MRI-current models?

The two analyses require similar preprocessing steps. Here's the list of the steps I've done and the respective scripts.

WP2A: I start from a dataset containing 22 folders (one per participant), each containing a T1 and a T2 scan (except for subject 16 who has only a T1).

  1. Renaming of anatomical scans: this renames the anatomical scans of each participant (i.e. sub-01_T1.nii etc).
  2. ROAST simulations: this script runs the ROAST simulations. In brief, ROAST outputs the following scans for each subject, while also using SPM routines for tissue segmentation: Voltage ("subjName_simulationTag_v.nii", unit in mV), E-field ("subjName_simulationTag_e.nii", unit in V/m) and E-field magnitude ("subjName_simulationTag_emag.nii", unit in V/m). The settings I have used for the simulation are: (t1, {'C3',1.0,'Fp2',-1.0},'T2', t2,'electype', 'pad', 'elecsize', [50 50 3], 'capType', '1020').
  3. Post ROAST preprocessing: ROAST outputs are in the ROAST model space. This script moves the results back to the MRI space, coregisters and normalises the electric field maps generated by ROAST. The script also normalise the T1 scan and all the masks.
  4. Ep values extraction from PEB result (Day-1 only): this script, starting from this .mat structure containing 66 PEBs (1 per participant / polarity), extracts the Ep values for each participant. The resulting file contains 66 matrices (participant 1 anodal, cathodal and sham, participant 2 ... 22).
  5. Estimation of posterior probability associated to each PEB extracted above. The script runs bayesian model averages for each PEB using the DCM function spm_dcm_peb_bmc. Results are saved in this .mat structure and used later on in the analyses to exclude connections with a posterior probability lower than 75%.
  6. WP2a e-magnitude measures estimation and correlation analysis. Steps:
    1. Load MNI template and M1/Th ROIs.
    2. Load .mat structure with Ep values and .mat structure with Pp values (Nb. Pp values are not used anymore);
    3. For each subject:
      1. Load normalised scan containing E-field magnitude (wsub-T1_emag.nii), normalised CSF, white and grey matter maps (wc1-2-3sub*.nii).
      2. Save DCM values related to the connections M1-M1, Th-Th, M1->Th and Th-> M1;
      3. Smooth E-field magnitude map using FWHM (4mm kernel);
      4. Mask E-field magnitude map with MNI template to exclude values outside the brain (useless if I then mask with CSF, wm and gm maps or with the M1/Th ROIs);
      5. Mask E-field magnitude map with M1 and Th ROIs and estimate means, medians and max electric-field values within the two ROIs;
      6. Save electric-field magnitude derived measures;
      7. Plot smoothed E-field magnitude map;
      8. Run 16 correlations: 4 DCM measures and 4 E-field measures (medians and max values).
      9. Plot correlations.

Questions:

  1. Electric field magnitudes or current densities?
  2. If so, how to deal with probabilistic masks?
  3. Should I threshold WM masks and apply binary erosion to remove the overlap between WM and GM?
  4. How to deal with Ep values which corresponding Pp is lower than our threshold (75%?)
  5. Should I mask out CSF tissue? Should I use a binary map containing only WM and GM?
  6. Hypotheses? Ideas?

Plots: Sticky note mind map - Sticky note mind map

References:

  1. Huang, Y., Datta, A., Bikson, M., & Parra, L. C. (2019). Realistic volumetric-approach to simulate transcranial electric stimulation—ROAST—a fully automated open-source pipeline. Journal of Neural Engineering, 16(5), 056006. https://doi.org/10.1088/1741-2552/ab208d
  2. Indahlastari, A., Albizu, A., Kraft, J. N., O’Shea, A., Nissim, N. R., Dunn, A. L., Carballo, D., Gordon, M. P., Taank, S., Kahn, A. T., Hernandez, C., Zucker, W. M., & Woods, A. J. (2021). Individualized tDCS modeling predicts functional connectivity changes within the working memory network in older adults. Brain Stimulation, 14(5), 1205–1215. https://doi.org/10.1016/j.brs.2021.08.003
  3. Albizu, A., Fang, R., Indahlastari, A., O’Shea, A., Stolte, S. E., See, K. B., Boutzoukas, E. M., Kraft, J. N., Nissim, N. R., & Woods, A. J. (2020). Machine learning and individual variability in electric field characteristics predict tDCS treatment response. Brain Stimulation, 13(6), 1753–1764. https://doi.org/10.1016/j.brs.2020.10.001
Owner
Davide Aloi
Doctoral Researcher at the University of Birmingham, UK. Centre for Human Brain Health. Investigating Disorders of Consciousness with fMRI and tDCS.
Davide Aloi
Non-stationary GP package written from scratch in PyTorch

NSGP-Torch Examples gpytorch model with skgpytorch # Import packages import torch from regdata import NonStat2D from gpytorch.kernels import RBFKernel

Zeel B Patel 1 Mar 06, 2022
Simulation-based performance analysis of server-less Blockchain-enabled Federated Learning

Blockchain-enabled Server-less Federated Learning Repository containing the files used to reproduce the results of the publication "Blockchain-enabled

Francesc Wilhelmi 9 Sep 27, 2022
PySLM Python Library for Selective Laser Melting and Additive Manufacturing

PySLM Python Library for Selective Laser Melting and Additive Manufacturing PySLM is a Python library for supporting development of input files used i

Dr Luke Parry 35 Dec 27, 2022
Graph Self-Attention Network for Learning Spatial-Temporal Interaction Representation in Autonomous Driving

GSAN Introduction Code for paper GSAN: Graph Self-Attention Network for Learning Spatial-Temporal Interaction Representation in Autonomous Driving, wh

YE Luyao 6 Oct 27, 2022
Bridging Composite and Real: Towards End-to-end Deep Image Matting

Bridging Composite and Real: Towards End-to-end Deep Image Matting Please note that the official repository of the paper Bridging Composite and Real:

Jizhizi_Li 30 Oct 31, 2022
Codes for AAAI22 paper "Learning to Solve Travelling Salesman Problem with Hardness-Adaptive Curriculum"

Paper For more details, please see our paper Learning to Solve Travelling Salesman Problem with Hardness-Adaptive Curriculum which has been accepted a

14 Sep 30, 2022
Source code of our BMVC 2021 paper: AniFormer: Data-driven 3D Animation with Transformer

AniFormer This is the PyTorch implementation of our BMVC 2021 paper AniFormer: Data-driven 3D Animation with Transformer. Haoyu Chen, Hao Tang, Nicu S

24 Nov 02, 2022
Official repository for the paper "Can You Learn an Algorithm? Generalizing from Easy to Hard Problems with Recurrent Networks"

Easy-To-Hard The official repository for the paper "Can You Learn an Algorithm? Generalizing from Easy to Hard Problems with Recurrent Networks". Gett

Avi Schwarzschild 52 Sep 08, 2022
[ ICCV 2021 Oral ] Our method can estimate camera poses and neural radiance fields jointly when the cameras are initialized at random poses in complex scenarios (outside-in scenes, even with less texture or intense noise )

GNeRF This repository contains official code for the ICCV 2021 paper: GNeRF: GAN-based Neural Radiance Field without Posed Camera. This implementation

Quan Meng 191 Dec 26, 2022
https://arxiv.org/abs/2102.11005

LogME LogME: Practical Assessment of Pre-trained Models for Transfer Learning How to use Just feed the features f and labels y to the function, and yo

THUML: Machine Learning Group @ THSS 149 Dec 19, 2022
Code for "AutoMTL: A Programming Framework for Automated Multi-Task Learning"

AutoMTL: A Programming Framework for Automated Multi-Task Learning This is the website for our paper "AutoMTL: A Programming Framework for Automated M

Ivy Zhang 40 Dec 04, 2022
The first public PyTorch implementation of Attentive Recurrent Comparators

arc-pytorch PyTorch implementation of Attentive Recurrent Comparators by Shyam et al. A blog explaining Attentive Recurrent Comparators Visualizing At

Sanyam Agarwal 150 Oct 14, 2022
OpenMMLab Image and Video Editing Toolbox

Introduction MMEditing is an open source image and video editing toolbox based on PyTorch. It is a part of the OpenMMLab project. The master branch wo

OpenMMLab 3.9k Jan 04, 2023
Semi-supervised semantic segmentation needs strong, varied perturbations

Semi-supervised semantic segmentation using CutMix and Colour Augmentation Implementations of our papers: Semi-supervised semantic segmentation needs

146 Dec 20, 2022
Pixel-Perfect Structure-from-Motion with Featuremetric Refinement (ICCV 2021, Oral)

Pixel-Perfect Structure-from-Motion (ICCV 2021 Oral) We introduce a framework that improves the accuracy of Structure-from-Motion by refining keypoint

Computer Vision and Geometry Lab 831 Dec 29, 2022
DeepLab is a state-of-art deep learning system for semantic image segmentation built on top of Caffe.

DeepLab Introduction DeepLab is a state-of-art deep learning system for semantic image segmentation built on top of Caffe. It combines densely-compute

Ali 234 Nov 14, 2022
Implementation of CoCa, Contrastive Captioners are Image-Text Foundation Models, in Pytorch

CoCa - Pytorch Implementation of CoCa, Contrastive Captioners are Image-Text Foundation Models, in Pytorch. They were able to elegantly fit in contras

Phil Wang 565 Dec 30, 2022
Worktory is a python library created with the single purpose of simplifying the inventory management of network automation scripts.

Worktory is a python library created with the single purpose of simplifying the inventory management of network automation scripts.

Renato Almeida de Oliveira 18 Aug 31, 2022
Art Project "Schrödinger's Game of Life"

Repo of the project "Team Creative Quantum AI: Schrödinger's Game of Life" Installation new conda env: conda create --name qcml python=3.8 conda activ

ℍ◮ℕℕ◭ℍ ℝ∈ᛔ∈ℝ 2 Sep 15, 2022
Moment-DETR code and QVHighlights dataset

Moment-DETR QVHighlights: Detecting Moments and Highlights in Videos via Natural Language Queries Jie Lei, Tamara L. Berg, Mohit Bansal For dataset de

Jie Lei 雷杰 133 Dec 22, 2022