BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:BMI Progress Reports 2021 // Prof. Herzog's Lab : D. Gordillo Lope
 z - The EEG multiverse of schizophrenia
DTSTART:20210120T121500
DTEND:20210120T130000
DTSTAMP:20260929T040036Z
UID:0f7ec31d455de659f0aa1e7c0a3296b2405ad3c6a357b37070f25d78
CATEGORIES:Conferences - Seminars
DESCRIPTION:D. Gordillo Lopez\nResearch on schizophrenia typically focuses
  on one paradigm\, for which clear-cut abnormalities between patients and 
 controls are established. Great care is taken to understand the underlying
  genetical\, neurophysiological\, and cognitive mechanisms\, which eventua
 lly may explain the clinical outcome. This approach has led to many import
 ant hypotheses about the causes of schizophrenia. One tacit assumption of 
 these deep rooting approaches is that paradigms tap into common and repres
 entative characteristics of the disease. Here\, we analyzed resting-state 
 electroencephalogram (EEG) of 121 schizophrenia patients and 75 age-matche
 d controls\, from which we extracted 194 features. Sixty-nine of these fea
 tures showed significant group differences with medium to large effect siz
 es\, indicating important abnormalities. To understand to what extent thes
 e features tap into the same aspects of the disorder\, we computed both Pe
 arson and partial least squares correlations. Surprisingly\, correlations 
 were very low\, except for very similar features\, where correlations were
  high\, suggesting that most features are sensitive to different abnormali
 ties. Using partial least squares regression\, we show that combining the 
 69 features increases predictability of clinical outcomes by more than a f
 actor of 5. We propose that complementing deep with shallow rooting approa
 ches\, where many roughly independent features are extracted from one para
 digm (or several paradigms)\, will strongly improve diagnosis and potentia
 l treatment of schizophrenia.\n 
LOCATION:Online
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR