GenMind™
Schizophrenia Research
The impact of smartphones and video game on real-time monitoring and schizophrenia treatment.
Co-authored by Michel Birnbaum and researchers from Johns Hopkins and Kyoto University.
Why It Matters: Negative symptoms and cognitive impairment in schizophrenia is different in every situation, showing up different at home than in a clinic. Thus, continuous, real-world monitoring can identify changes that scheduled appointments can miss.
Nature Schizophrenia
Using machine learning to identify schizophrenia and depression via objective linguistic, speech, facial, and motor behavorial cues.
Co-authored by Michel Birnbaum and researchers from Nanyang Technological University and National University of Singapore.
Why It Matters: This study provides clear evidence that objective, behavorial signals leveraged by machine-learning models can help diagnose, assess, and monitor patients with schizophrenia and depression.
Springer Singapore
Using linguistic cues to identify how patients with schizophrenia speak, and what it reveals about their condition.
Co-authored by Michel Birnbaum and researchers from Nanyang Technological University.
Why It Matters: This study explores the method of using a monitoring tool to pick up word-choice patterns passively from natural conversation, removing the need of a structured clinical setting.
IEEE
Analyzing voice and speech patterns in an automated objective pipeline to provide insight into symptoms of schizophrenia and depression.
Co-authored by Michel Birnbaum and researchers from Nanyang Technological University.
Why It Matters: By providing an automated method to analyze symptoms of these two common mental disorders, it suggests that these kinds of tools can support ongoing tracking outside the clinical visit.
IEEE
Using automatic transcriptions of interviews to observe negative symptoms in schizophrenic patients.
Co-authored by Michel Birnbaum and researchers from Nanyang Technological University.
Why It Matters: This automated method is a practical step towards scaling this type of assessment in a a clinical setting. The study is also showcases the benefit of using lexical features in the understanding and diagnosis of schizophrenia.
IEEE
Using objective conversational cues in tracking negative symptoms in schizophrenia.
Co-authored by Michel Birnbaum and researchers from Nanyang Technological University.
Why It Matters: Automated symptom analysis reduces the reliance of clinical availability, expanding where and how often assessments can be done.
NeuroBrowser™
International Journal of Neural Systems
Proposing a patient-independent seizure detector to automatically detect seizures in both scalp EEG and intracranial EEG (iEEG).
Co-authored by Michel Birnbaum and researchers from Nanyang Technological University, KU Leuven, and McGill University.
Why It Matters: The proposed seizure detector can detect seizures in adult EEGs in less than 15s for a 30min EEG. This can aid clinicians in reliably expediting seizure identification, providing more time for devising proper treatment.
Journal of Neural Engineering
Analyzing epileptic EEGs without IEDs
Co-authored by Michel Birnbaum and researchers from Nanyang Technological University, University of Pennsylvania, and McGill University.
Why It Matters: These results are a jumping off point for automated detection of epilepsy. It is one of the first to analyze epileptic EEGs without IEDs, thus opening up an under-explored option in epilepsy diagnosis.
IEEE
Proposing an artificial detection module to reject artifact segments, leading to a cleaner EEG for further analysis.
Co-authored by Justin Dauwels and researchers from Nanyang Technological University.
Why It Matters: Artifacts cause EEG misinterpretation. The proposed system can reject a substantial amount of artifacts while only removing a small fraction of a clean EEG, improving readability of recordings.
International Journal of Neural Systems
Evaluating features that may provide reliable IED detection and EEG classification.
Co-authored by Justin Dauwels and researchers from Nanyang Technological University, Stanford University, and Massachusetts General Hospital.
Why It Matters: The proposed classification system in this study only takes a few seconds to analyze a 30-min routine EEG. This may help reduce the human effort required for epilepsy diagnosis.
International Journal of Neural Systems
Proposing three automated approaches to detect slowing in EEG.
Co-authored by Justin Dauwels and researchers from Nanyang Technological University.
Why It Matters: These automated approaches achieved results comparable to human experts. Additionally, the deep-learning-based system can process a 30 minute EEG in 4 seconds, and can be deployed to assist clinicians in interpreting EEGs.
International Journal of Neural Systems
Classifying EEG as epileptic or normal based on multiple modalities extracted from the interictal EEG.
Co-authored by Justin Dauwels and researchers from Nanyang Technological University, Stanford University, and Massachusetts General Hospital.
Why It Matters: The six-center study shows strong generalization of effectiveness. The system can also process routine EEGs in seconds, effectively aiding clinicians in diagnosing epilepsy.
Our Awards and Certifications
MedTech Innovator - Asia Pacific 2019
AI NY Summit 2022 Runner Up
