Independent research · 2025–26
Neuronal cell-type classification
A study of whether morphological and electrophysiological features can help distinguish neuronal cell types across human and mouse datasets.
- Timeline
- June 2025–March 2026
- Data
- Allen Brain Atlas
- Type
- Independent research
- Recognition
- 2026 · First Place Trophy · Grades 9–12
The question
What information is useful?
Neurons can be described by their shape and by how they respond electrically. I wanted to understand how much each type of information contributes to classification.
The work
Preparing comparable data
I prepared human and mouse datasets and aligned morphological, electrophysiological, and structural-layer features for analysis.
Comparing models
I compared decision trees, random forests, and LSTMs rather than assuming that the most complex model would be the most informative.
Looking at errors
Cross-validation and confusion matrices helped show which categories were consistently distinguishable and where the models confused related cell types.
Recognition
Washington State Science & Engineering Fair
The neuronal classification project received the fair’s 2026 First Place Trophy for grades 9–12.