A central principle of my research is that scientific progress in data-driven neuroscience and digital health depends on openness, reproducibility, and scalable infrastructure. Alongside theoretical and applied research, I have led and contributed to the development of open platforms, datasets, and software tools that support collaborative research, transparent benchmarking, and real-world translation in epilepsy and neural engineering.
Research Platforms and Ecosystems
Epilepsyecosystem.org
Epilepsyecosystem.org is a large-scale, cloud-based research platform designed to support reproducible and collaborative epilepsy research. It enables researchers to analyse ultra-long-term intracranial EEG data, benchmark seizure prediction algorithms, and share results using standardized pipelines.
The platform was developed to address long-standing challenges in seizure prediction research, including limited data access, inconsistent evaluation protocols, and poor reproducibility across studies. By providing a shared infrastructure and common evaluation framework, Epilepsyecosystem.org facilitates transparent comparison of methods and accelerates progress toward clinically viable seizure forecasting systems.
The ecosystem continues to expand to incorporate additional data modalities, methods, and international collaborators.
Melbourne University–AES–MathWorks–NIH Seizure Prediction Challenge
This international data science challenge was organised to catalyse innovation in seizure prediction by bringing together researchers from neuroscience, engineering, and machine learning communities. Hosted on Kaggle, the challenge provided access to large intracranial EEG datasets and a standardized evaluation framework.
The challenge demonstrated the value of open competitions for advancing algorithmic performance, identifying robust features, and establishing reproducible benchmarks. Insights and methodologies emerging from this initiative have directly informed subsequent research and platform development within the Epilepsyecosystem.
Software Tools
Seer.py
Seer.py is an open-source Python library developed to support data access, analysis, and reproducible workflows within the Epilepsyecosystem. It provides programmatic interfaces for working with large-scale EEG datasets and associated metadata, enabling researchers to integrate ecosystem data into their own analysis pipelines.
The library is designed with scalability and transparency in mind, allowing methods to be shared, inspected, and reproduced across institutions. Seer.py supports both exploratory research and structured benchmarking experiments.
Source code and documentation are available via the project’s public repository.
Data Philosophy and Reproducibility
Across all software and data initiatives, a consistent design philosophy is followed:
• Emphasis on reproducibility, with standardized data formats and evaluation protocols
• Support for large-scale, continuous recordings, reflecting real clinical conditions
• Enablement of collaborative research, rather than isolated algorithm development
• Clear separation between data access, method implementation, and evaluation
This approach ensures that research outcomes are robust, interpretable, and transferable beyond individual studies.
Using These Resources
Researchers interested in using the platforms or software tools described above are encouraged to explore the available documentation and published studies linked to each resource. For collaborative projects, data access questions, or integration into new research initiatives, direct enquiries are welcome.
Citation and Acknowledgement
If you use data, software, or infrastructure associated with these platforms in your research, please cite the relevant publications and acknowledge the corresponding resources. Proper citation helps sustain open research ecosystems and supports continued development.