One of my key interests is enhancing research and clinical practice in clinical neuroscience. Between 2014 and 2018, I have focused on EEG-based diagnosis in disorders of consciousness (DoC). My interdisciplinary research activities have led to software solutions for automated large-scale processing and reporting of clinical EEG, introduced at the neuroscience workshop of the International Conference on Machine Learning (ICML) 2015 and leading to the AutoReject algorithm. Ever since, these tools have been used at the Pitié-Salpêtrière hospital, Paris, France and have led to several original research contributions and critical commentaries that I have co-authored.
My work on DoC progressively developed into the, so far (2018), most extensive validation of EEG-biomarkers of consciousness based on more than 320 clinical EEG-recordings from two top-notch European clinical research groups (Paris, Pitié-Salêtrière, Paris, France, and, Coma Science Groups, Liège, Belgium). Our key paper demonstrates the possibility of cross-protocol and cross-site generalization of predictions of diagnosis obtained from machine learning algorithms. If you have no time to read the paper, this twitter thread gives you the gist:
1/ New paper out! @Brain1878 https://t.co/l3VIfg4s9F Robust #MachineLearning of diagnosis from #EEG in disorders of #consciousness powered by @CEAParisSaclay @cdf1530 @Coma_Science @dc_uba @ERC_Research @Inria_Saclay @Inserm @mne_news @NeuroSpin_91 @icm_institute @scikit_learn pic.twitter.com/JsDMUYWw6C— Denis A. Engemann (@dngman) October 4, 2018