- #Picard principal definition how to
- #Picard principal definition code
The requirements to run the python scripts can also found in the scripts folder in the requirements.txt file.
building_model.py : script to compute the regression model, the permutation test and the bootstrap test. prepping_data.py : script to load the fMRI data in Nifti-like format and to extract the signal according to a nilearn pre-defined mask or to a pre-computed mask. data_to_json.py : script to save the fmri data (predictive variable), the target variable (predicted variable) and the group variable (for non iid data: multiple observations per participant, optional) in json format. #Picard principal definition code
This script also includes code to save the train and test sets (in numpy format), the coefficients of the regression model (in nii.gz format), the model (in pickle format), the output of the permutation test (in json format) and the bootstrap test (in pickle format).
main.py : main script calling functions in prepping_data.py and building_model.py. Deliverablesįour python scripts can be found in the scripts folder: All the deliverables were completed and can be found on this repository (see the Deliverables section below). The final presentation of the project was delivered on 20th August 2021. This project was initiated by Marie-Eve Picard on 2nd August 2021 as part of the PSY6983 course. Github repository containing the scripts, graphs, report and relevant information. MarkDown README.md documenting all important information about the project. Static and interactive visualizations of the results. Illustration of the experimental paradigme.ĭistribution of the FACS scores across the datasetsīy the end of this course, I will provide Each participant completed 2 runs of 8 painful trials, resulting to a total of 797 observations after removing data with movement artefacs (dataset 1: n = 533 observations dataset 2: n = 264 observations). Their facial expression was recorded during this time by the use of a MRI-compatible camera. Moderately painful stimuli were applied while they where in the MRI scanner. It includes 55 participants (women = 28, men = 27) aged between 18 and 53 years old. The dataset that will be used for this project is a secondary private dataset (not open access yet). The python scripts and figures will be added to the github repository. The analysis will be run on compute canada. Matplotlib, seaborn, plotly: to plot some figures. Nilearn: to extract and to visualize the fMRI data. Scikit-learn: for machine learning stuff. Nibabel: to load the fMRI contrast images (from. Pandas: to manipulate the behavioral dataframe. Python scripts to write the code for the analysis. The completion of this project will require the following tool: #Picard principal definition how to
Learn how to run analysis on compute canada and don't be afraid of it ✔️. Apply machine learning to neuroimaging data ✔️. We can compute a composite score (FACS score) taking into account the frequency and the intensity of the contraction of these action units. The facial expression can be measured with the Facial Action Coding System (FACS, Ekman & Friesen, 1978), which is based on the possible action of muscles or groups of muscles, called action units. Facial movements related to the experience of pain are associated not only with the sensory dimension of pain (i.e., pain intensity) but also with its affective dimension (i.e., the unpleasantness of pain). Project definition Backgroundįacial expression of pain is an important non-verbal informative element of communication, signaling an immediate threat and a need for help. I'm also interested in the different methods used to analyze neuroimaging data. 
The goal of my master project is to predict the occurrence and intensity of the facial expression of pain from brain activity (fMRI data) in response to a tonic painful stimuli. I am interested in the neural correlates of pain and the communication of pain.
Currently a Master's student in Psychology at Université de Montréal. Predicting facial expression of pain in healthy adults from fMRI data
0 kommentar(er)
