Uploaded June 2024 | Updated September 2026, 1 hour ago
While machine learning has become very popular for neuroimaging data, it remains non-trivial to use this set of techniques properly. It can indeed be challenging and time-consuming not only to implement machine learning models for fNIRS data classification, but also to evaluate such models in a way that reflects the performance on unseen data (generalisation capabilities). This is key however in order to estimate realistically the capabilities of real-world brain-computer interfaces for example.
In this webinar, our speaker introduces some of the challenges of machine learning with fNIRS data, and shows how the BenchNIRS Python framework can be used to simplify the implementation, fine-tuning and rigorous evaluation of machine learning models for fNIRS data classification. Furthermore, Johann presents a case study of how BenchNIRS was used to initiate the largest benchmarking of machine learning on existing open-access fNIRS datasets, comparing popular models on various tasks and paradigms.
While machine learning has become very popular for neuroimaging data, it remains non-trivial to use this set of techniques properly. It can indeed be challenging and time-consuming not only to implement machine learning models for fNIRS data classification, but also to evaluate such models in a way that reflects the performance on unseen data (generalisation capabilities). This is key however in order to estimate realistically the capabilities of real-world brain-computer interfaces for example.
In this webinar, our speaker introduces some of the challenges of machine learning with fNIRS data, and shows how the BenchNIRS Python framework can be used to simplify the implementation, fine-tuning and rigorous evaluation of machine learning models for fNIRS data classification. Furthermore, Johann presents a case study of how BenchNIRS was used to initiate the largest benchmarking of machine learning on existing open-access fNIRS datasets, comparing popular models on various tasks and paradigms.
