Naslov
Big data approach to assessment of soldier stress resilience

Autori
Ćosić, Krešimir ; Popović, Siniša ; Kovač, Bernard

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
1st International Congress of the International College of Person-centered Medicine: Whole Person in Health Education and Trainig : [book of Abstracts] / - , 2013, X-x

Skup
Knjige na Google Playu 1st International Congress of the International College of Person-centered Medicine: Whole Person in Health Education and Trainig

Mjesto i datum
Zagreb, Hrvatska, 7-10.11.2013

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Podatak o recenziji nije dostupan

Ključne riječi
big data ; stress resilience ; soldier

Sažetak
The concept of new multidimensional and multimodal metrics for assessment of soldier stress resilience is presented, which is based on comprehensive analysis of a few hundred physiological, acoustic, facial and EEG features computed during soldier elicitation with specific multimodal stimulation using mission-relevant affective databases. Augmentation of this dynamic data streams with resilience-relevant biomarkers, like: COMT Val158Met polymorphism, catabolic regulation of synaptic dopamine, immune/inflammatory endocrine and autonomic response, as well as with fMRI/DTI images of key brain circuitry and their connectomics, including a variety of questionnaires like subjective self- reports, personality traits, phenotypes etc. makes the problem of soldier resilience assessment much more complex. Aggregation and cross-correlation longitudinal analysis of these unstructured multidimensional and multimodal datasets along soldier life cycle, including development of soldiers personalized predictive models, leads to real big data problem. Crunching, learning and reasoning on such huge biological, neural, behavioral and environmental personalized big datasets require multidisciplinary and interdisciplinary expert teams. In order to understand complex, nonlinear, and stochastic interactions among soldier’s environment, behavior and biology, we must move from group-based analysis toward individuals personalized big data analysis based on sophisticated machine learning mathematical methods and techniques which can extract valuable scientific and applicable results and knowledge.

Izvorni jezik
Engleski

Znanstvena područja
Elektrotehnika, Računarstvo, Psihologija

POVEZANOST RADA