Grant: Master's Award: Canada Graduate Scholarships
Université de Montréal
$17,500.00
Consciousness and artificial intelligence: using the brain's response to anesthesia and machine learning to predict consciousness recovery in coma and disorders of consciousness
FY 2022-23
Summary
Advanced survival and life-sustaining technologies have led to an increased prevalence of brain-injured individuals in disorders of consciousness (DOC), who have minimum to no behavioural responsiveness and unknown levels of awareness. Establishing an accurate prognosis in acute DOC is a fundamental shortcoming of clinical practice. Our team has recently developed an accurate prognostic index for DOC – the Adaptive Reconfiguration Index (ARI) – which quantifies the reconfiguration of brain activity in response to propofol anesthesia, using electroencephalography (EEG). In a preliminary study, the ARI detected with 100% accuracy the return of consciousness within 3 months in 10 DOC patients. Building upon these results, this study aims to validate the accuracy of the ARI for the prognostication of consciousness recovery in coma and DOC, and use machine-learning to improve the ARI and its prognostic capacity. To calculate the ARI, we will record the brain’s response to anesthesia in 30 non-sedated patients in a DOC. We will investigate whether the ARI can predict recovery of consciousness at 3-month follow-up and assess the ARI’s prognostic performance for consciousness recovery. In parallel, we will train machine learning models for the same purpose using additional EEG features of 20 patients. The most performant model’s prognostic accuracy will be tested on the remaining 10 patients. Upon completion of this study, we expect to have developed and validated a clinically relevant tool for the prognosis of individuals in acute DOC, which will immediately be available for clinical translation. Improving DOC prognosis will meaningfully contribute to clinical management and decision making and will lead to better-adapted rehabilitative strategies and better patient outcomes.
Department
Program
Master's Award: Canada Graduate Scholarships
Recipient
Location
Quebec, Canada
Recipient Type
Academia
Type
Grant
Source
CIHR Funding Decisions Database
As published
Id
485931
Name
Martinez Villar Guillermo E
Region
Québec
Country
Canada
Orgtype
Institution Paid; Research Institution
Prcname2
Special Cases - Awards Programs
Projecttitle
Consciousness and artificial intelligence: using the brain's response to anesthesia and machine learning to predict consciousness recovery in coma and disorders of consciousness
Programtype2
Studentship Programs
Orgnameinp2
Université de Montréal
Cihramount2
$17,500
Cihrequipment2
$0
Programname2
Master's Award: Canada Graduate Scholarships
Keyworddelim
Anesthesia; Classification; Coma; Consciousness; Electroencephalography; Functional Connectivity; Machine Learning; Prognosis
Abstract
Advanced survival and life-sustaining technologies have led to an increased prevalence of brain-injured individuals in disorders of consciousness (DOC), who have minimum to no behavioural responsiveness and unknown levels of awareness. Establishing an accurate prognosis in acute DOC is a fundamental shortcoming of clinical practice. Our team has recently developed an accurate prognostic index for DOC – the Adaptive Reconfiguration Index (ARI) – which quantifies the reconfiguration of brain activity in response to propofol anesthesia, using electroencephalography (EEG). In a preliminary study, the ARI detected with 100% accuracy the return of consciousness within 3 months in 10 DOC patients. Building upon these results, this study aims to validate the accuracy of the ARI for the prognostication of consciousness recovery in coma and DOC, and use machine-learning to improve the ARI and its prognostic capacity. To calculate the ARI, we will record the brain’s response to anesthesia in 30 non-sedated patients in a DOC. We will investigate whether the ARI can predict recovery of consciousness at 3-month follow-up and assess the ARI’s prognostic performance for consciousness recovery. In parallel, we will train machine learning models for the same purpose using additional EEG features of 20 patients. The most performant model’s prognostic accuracy will be tested on the remaining 10 patients. Upon completion of this study, we expect to have developed and validated a clinically relevant tool for the prognosis of individuals in acute DOC, which will immediately be available for clinical translation. Improving DOC prognosis will meaningfully contribute to clinical management and decision making and will lead to better-adapted rehabilitative strategies and better patient outcomes.
Competitiondate
202212
Approvedterm2
1 yr 0 mth
Pinamesdelim
Martinez Villar, Guillermo E
Approvedterm
1 an 0 mois
Cihrcontribution2
~ $17,500~ $0~