CL202INTERMEDIATECLINICAL

AI for Neurology

The flagship clinical course: machine learning across neurological signals, imaging, movement and clinical records.

Course Information

Course Information
ModeInstructor-led · live lecture and guided lab
TermFA1 · Aug–Sep · proposed
Duration8 weeks · two 90-minute sessions weekly · 24 live hours, plus independent work
Credits1.5 credits
PrerequisitesNo prior AI course required. Week 1 introduces machine learning; CS101 or equivalent Python preparation is recommended. CL101 is helpful.
Who it is forNeurologists, neurology residents, medical students, neurophysiology technologists, rehabilitation professionals and engineers.
FacultyTo be announced
AssessmentWeekly quizzes 20% · two case assignments 40% · clinical mini project 40%. 60% to pass.

About the course

Each clinical question is examined through data, modelling, validation and workflow. Guided notebooks use public datasets and pretrained models. Learners distinguish a promising prediction from evidence of clinical benefit.

Learning outcomes

  • Match neurological data to suitable AI methods and evaluate sensitivity, predictive values, calibration and thresholds
  • Build a seizure-detection pipeline with patient-wise validation; assess stroke imaging, tumour segmentation and digital biomarkers
  • Evaluate language-model outputs and examine external validation, clinical usefulness and implementation limits

Course modules

8 modules

  • Clinical questions, supervised learning, confusion matrices and calibration

  • first classifier on tabular data.

Labs and projects: Weeks 5–8, teams frame a clinical question, analyse a public dataset, evaluate errors and thresholds, and submit a report and ten-minute presentation. This course project is separate from the RP301 certificate capstone.

Books and readings: Datasets: CHB-MIT, TUH EEG, RSNA head CT, BraTS, ISLES, Sleep-EDF and Parkinson’s datasets; some require registration. Readings: Topol (2019), High-performance medicine; TRIPOD+AI, CLAIM and ICMR AI ethics guidelines (2023).