This repository contains portfolio assignments and descriptions of the assignments (as pdf files), which were done during the experimental methods II course as part of Cognitive Science BA´s degree at Aarhus University
2019 spring semester
The course consists of two main parts: During the first part of the course, students are introduced to how the general linear model is used in the analysis of brain scan data, for example those deducted from an fMRI and EEG/MEG scan. The range of methods of analysis is increased in the second part of the course to include multivariate analyses.
File Portfolio assignments.pdf contains a full description of the assignment instructions. File Portfolio_exam.pdf is a compilation of the solutions to all of the portfolio assignments.
| Assignment | Description |
|---|---|
| Portfolio 1: Regression | Sleep deprivation excercise where we look at response time as a function of days of sleep deprivation and practice linear regression analysis method |
| Portfolio 2: Vectors and matrices | In this excercise we are looking at response time as a function of days of sleep deprivation using matrices. It consists of 4 parts: Linear regression, Images and matrices, Brains and matrices, Two equations with two unknowns |
| Portfolio 3: fMRI regression | fMRI single voxel data excercise. We analyse a time-series from a single voxel in auditory cortex (transverse temporal gyrus, MNI coordinate: [-46,-20,6], with all time-points converted into a vector). The task is to perform a regresion with hemodynamic response to the different story types (fiction and factual) as independent variables using different models and also adding an additional covariate |
| Portfolio 4: fMRI preprocessing | In this exercise we are preparing fMRI data for analysis and look at some of the output. We look at all the fMRI data from one participant |
| Portfolio 5: Model specification | In this exercise, we make a model for the analysis of the data that we preprocessed in Portfolio 4. The model is specified by onsets for the different stories and by their duration, the onsets for emotional ratings of the stories |
| Portfolio 6: fMRI model estimation | In this exercise, we estimate the model of the data that we designed in Portfolio 4. We get the results and report them, if there are any. An SPM software package (Matlab) was used to obtain the results. For comparison of different conditions we employ T-contrast and F-contrast |
| Portfolio 7: Mixed effects | We practice mixed effects models (including 2-way and 3-way interactions) using data from two experimental studies: Emotional faces experiment & Tryptophan depletion study analysis |
| Portfolio 8: EEG Stop-Signal-NoGo | Reporting results of an EEG data after conducting the preprocessing |
| Portfolio 9: fMRI group analysis | We investigate the fMRI brain scans from emotional face experiment which consists of the following parts: Preprocessing and modeling of individual participants, Investigating analysis of a single participant, Group analysis |
| Portfolio 10: PCA factor analysis | Real World Immitating Task employing Principal Component Analysis and Factor analysis |
Knowledge: After completing the course, students will have gained knowledge of:
- How the general linear model (GLM) is used for the analysis of brain scan data
- Basic assumptions about multivariate statistics
- Various types of multivariate analyses, such as principal component analysis, canonical correlation and linear discriminant analysis
Skills: After completing the course, students will be able to:
- Understand, develop and carry out complex experiments
- Assist in analyses of brain scan data that use the general linear model
- Analyse multiple types of data using multivariate statistics
Competences: After completing the course, students will be able to:
- Explain what multivariate analysis includes and when this kind of analysis is suitable
- Compare univariate and multivariate statistics and identify strengths and weaknesses of the two methods
- Identify the potential for further learning in relation to analyses of brain data using the general linear model
- Identify the potential for further learning in relation to multivariate analyses