BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Challenges of building clinical biomarkers from M/EEG: multimodal 
 modeling with missing data and robust regression on power spectra - Denis 
 Engemann\, Parietal\, INRIA
DTSTART:20200622T113000Z
DTEND:20200622T130000Z
UID:TALK148786@talks.cam.ac.uk
CONTACT:63079
DESCRIPTION:In clinical neuroscience\, success often depends on reading ou
 t multiple modalities\, i.e.\, brain images and physiological signals. How
 ever\, clinical reality often sets limits on data availability. Is combini
 ng multiple modalities for predictive modeling worth the extra effort when
  data is regularly incomplete? In [1]\, we proposed a multi-modal machine 
 learning model with explicit support for handling missing modalities. Comb
 ining MRI\, fMRI and magnetoencephalography on the Cam-CAN database not on
 ly significantly enhanced age prediction but also facilitated detection of
  age-related cognitive decline captured by the estimated brain age delta. 
 In\, particular\, combining MEG with MRI yielded enhanced detection of cha
 nges in fluid intelligence\, sleep quality and memory function\, highlight
 ing the complementarity of these distinct biomedical signals. Strikingly\,
  the added value of MEG was best explained by relatively simple features\,
  i.e.\, the spatial distribution of fast brain rhythms in the beta/alpha r
 ange. These results potentially open the door to clinical translation via 
 EEG-technology that is widely available in the hospital setting.\nUnfortun
 ately\, MRI scans are not always available\, closing the door to source mo
 deling with individual anatomy. What then? Call linear models for rescue? 
 While very effective for regressing biomedical outcomes on M/EEG signals\,
  they fail systematically if the cortical generator of an observed behavio
 r is oscillatory. In that case\, volume conduction induces distortions on 
 extracranial signals mitigating the applicability of linear models. Howeve
 r\, accurate modeling volume conduction depends on the availability of ind
 ividual MRIs in the first place. In [2\,3] we demonstrate through mathemat
 ical analysis\, simulations and prediction of age from MEG (Cam-CAN) and E
 EG (Temple University Hospital) how to\, nevertheless\, construct predicti
 ve linear models in different data generating scenarios. We conclude that 
 Riemannian geometry offers a practical alternative to source localization 
 when predicting from power spectra\, potentially enabling end-to-end learn
 ing without preprocessing.\n\nReferences\n\n[1] Engemann\, DA.\, Kozynets\
 , O.\, Sabbagh\, D.\, Lemaitre\, G.\, Varoquaux\, G.\, Liem\, F.\, & Gramf
 ort\, A. (2020). Combining magnetoencephalography with MRI enhances learni
 ng of surrogate-biomarkers. eLife. 9:e54055. 10.7554/eLife.54055\n\n[2] Sa
 bbagh\, D.\, Ablin\, P.\, Varoquaux\, G.\, Gramfort\, A.\, & Engemann\, DA
 . (2019). Manifold-regression to predict from MEG/EEG brain signals withou
 t source modeling. In H. Wallach\, H. Larochelle\, A. Beygelzimer\, F. Alc
 hé-Buc\, E. Fox\, & R. Garnett (Eds.)\, Advances in Neural Information Pr
 ocessing Systems 32 (pp. 7321–7332).\n \n[3] Sabbagh\, D.\, Ablin\, P.\,
  Varoquaux\, G.\, Gramfort\, A.\, & Engemann\, DA. (2020   ). Predictive r
 egression modeling with MEG/EEG: from source power to signals and cognitiv
 e states. NeuroImage. https://doi.org/10.1016/j.neuroimage.2020.116893
LOCATION:Zoom (contact mri.admin@mrc-cbu.cam.ac.uk for attendance details)
END:VEVENT
END:VCALENDAR