Patrick Filima

Patrick Filima

Research Software Engineer

Computational Neuroscientist

© 2026 Patrick Filima

Research Programs & Studies

Computational neuroimaging, brain morphometry, and clinical AI validation

My research combines computational neuroscience with robust software engineering to study brain morphology, disease biomarkers, and medical imaging AI.

By investigating neurodegenerative changes in under-represented African populations and developing reproducible imaging workflows, my research aims to make neuroimaging diagnostics more equitable and methodologically sound.

Domains

Core Research Themes

Computational Neuroimaging

Structural and functional MRI analysis, brain morphometry, volumetric profiling, and automated segmentation pipelines for neurodegenerative conditions.

Key Methods:
Structural & functional MRI
Subcortical nuclei segmentation
Cortical thickness modeling
Artifact correction & QC

Medical AI & Robustness

Developing and validating explainable deep learning models on real-world clinical scans, with an emphasis on low-field robustness and calibration.

Key Methods:
MONAI & PyTorch architectures
Attribution stability (Grad-CAM, SHAP)
Calibration error (ECE)
Domain generalization

FAIR Research Infrastructure

Designing open data ecosystems, BIDS-compliant curation pipelines, and reproducible scientific software for neuroscience research in low-resource settings.

Key Methods:
BIDS dataset curation
FAIR metadata schemas
Dockerized analysis pipelines
Open-access data governance
Active Investigations

Ongoing Clinical & Methodological Studies

Published Work
Neuroimaging & MorphometryActive Study

Asymmetry in Thalamic Gray Matter Changes in Nigerian Parkinson's Disease Patients

Investigating asymmetrical patterns of thalamic gray matter alterations in Nigerian Parkinson's Disease patients using automated volumetric MRI segmentation and statistical shape modeling.

Objectives:
  • Map hemispheric differences and volumetric atrophy across disease progression stages
  • Correlate regional structural alterations with motor and cognitive clinical scores
  • Characterize subcortical biomarkers specific to under-represented African cohorts
Methodology:

Structural 1.5T/3T MRI, FSL FIRST segmentation, FreeSurfer subcortical pipeline, GLM statistical modeling.

Expected Scientific Impact:

Provides essential population-specific baseline data for neurodegenerative diagnostic algorithms in West African populations.

Subcortical MorphometryActive Study

Volumetric Assessment of Individual Thalamic Nuclei in Nigerian Parkinson's Cohorts

High-resolution segmentation of discrete thalamic sub-nuclei (pulvinar, ventral intermediate, mediodorsal) to evaluate focal volume degradation and disease severity correlations.

Objectives:
  • Quantify volume changes in specific thalamic sub-nuclei
  • Establish correlations between nucleus volume loss and disease progression
  • Benchmark regional measurements against international normative reference datasets
Methodology:

Automated sub-nuclear parcellation, Bayesian segmentation algorithms, regression modeling.

Expected Scientific Impact:

Identifies candidate structural markers for early-stage neurodegenerative differentiation.

Behavioral & Social NeuroscienceActive Study

Cross-Cultural Behavioral Analysis of Screen Time and Social Connectivity

A cross-cultural study evaluating the impact of digital screen time on interpersonal relationships, sleep quality, and psychological well-being between German and Nigerian cohorts.

Objectives:
  • Analyze cross-cultural patterns of technology and mobile screen usage
  • Assess associations between digital engagement and self-reported social connectivity metrics
  • Identify cultural mediators influencing digital habit formation
Methodology:

Mixed-methods psychometric assessment, digital tracking metrics, multivariate statistical analysis.

Expected Scientific Impact:

Contributes cross-cultural data to international digital health and behavioral neuroscience frameworks.