A description of my research project and goals
The project is tentatively titled "Using Machine Learning
to Detect Changes in Working Memory Load and Executive
Functioning". The project will conduct a study where
participants perform the n-back task of varying difficulty
levels (n: 0-2) while a fNIRS device takes readings of the
hemodynamics in areas relevant to working memory. Data taken
from the study will be used to train a machine learning model;
this model will classify data into those which were taken while
performing the 0-back, 1-back, or 2-back task.
The final goal of the project is to develop a robust classifier
for developing a robust classifier that can classify working
memory demand at the 3 levels. This model will be implemented
in future studies with more authentic tasks and it will try to
classify the working memory demand, based on its training on
the n-back data.
As decided by my mentor and I, the project should follow this
- Review literature to understand what is known in
psychology and neuroscience about working memory demand
and executive functioning, with a focus on the "n-back"
task. Emphasis on understanding what the n-back task is
and why it is used in studies of working memory and
executive. Also emphasis on learning what neuroscientists
have discovered about the brain activation associated
with different levels of this task using fNIRS, fMRI and
PET brain imaging. (Weeks 1-2)
- Write software to enable study participants to
perform the task. It will need to integrate with all of
the lab's experimental software and will log all the data
from the user. (Weeks 2-4)
- Put together an experiment design and plan the
experiment (Weeks 3-5)
- Recruit participants for study (Week 5)
- Run study (Weeks 6-7)
- Machine learning analysis (Weeks 7-10)
- Gain proficiency with a new language (Python, C#,
- Be able to properly utilize an fNIRS system.
- Successfully train a model to classify working memory load.