Stroke Risk Patient Data With CT Scans

A multimodal dataset for stroke risk prediction contains demographic, medical history, and lifestyle data alongside CT scan images of normal and stroke-affected conditions. The dataset is 5.5 KB in size and was published by author Ramya Sree K on figshare in April 2026. It was used to develop a federated learning model combining MLP and GRU architectures.

Use Cases

• Predict stroke risk from static features like demographic data, medical history, and lifestyle information using classification models.
• Analyze sequential features from CT scan image data to identify patterns associated with stroke conditions.
• Train multimodal models, such as MLP-GRU hybrids, where MLP processes static tabular data and GRU processes sequential imaging features.
• Benchmark federated learning frameworks for healthcare by splitting the multimodal data across simulated clients.

Strengths

• Dataset includes two complementary modalities: structured tabular data and CT scan images.
• Data preprocessing included imputation, normalization, and handling of class imbalance with SMOTE.

Limitations

• The dataset is very small at 5.5 KB, indicating extremely limited sample size.
• Specific row counts, column names, and detailed data schema are not provided.
• Geographic and demographic coverage of the patient data is unknown.

Provenance

Source

figshare

Collection Method

Collection method not specified; data was used for a federated learning research project.

Time Range

null

Freshness

null

Geography

null