Grant: Alliance Grants
University of Manitoba
$25,000.00
Autocorrelative and Multicollinearity Structures of Mammalian Cerebral Physiologic Data Streams for Future Application in Machine Learning Pipelines
FY 2023-24
Summary
Trajectory modelling of physiologic state (pre-emptively anticipating future states) is considered a 'holy grail' for various physiologic, data science and life/veterinary science fields. To facilitate this temporally resolved physiologic forecasting, modelling of the data is required which now often leverages machine learning/artificial intelligence driven processes. Typically, physiologic data, in current models, is inputted without accounting for inherent autocorrelative structures of parent signals or multicollinearity in vector space (the linear impact within interlinked data). Violating such priors of linearity leads to model errors and uncertainty in significance of forecasted values. Currently, our understanding of the time-series autocorrelative structures of cerebral physiology and the interconnected multicollinearity between them is unknown. Such work is further impaired by the impact of data resolution and computational lag time required for optimal physiological model creation. To improve physiological models and forecasting such knowledge gaps need to be fulfilled. Leveraging existing globally unique high-fidelity, high-frequency multi-modal cerebral physiologic data sets in humans from both the University of Manitoba (UM) and University of Helsinki (UH), this NSERC Alliance International Catalyst project aims to apply time-series methodologies (UM) and leverage collaborative expertise in their consideration for machine learning pipeline development (UM and UH), characterizing: the autocorrelative structure/multicollinearity features, the impact of data resolution, and computational considerations with derivation of models to account for such autocorrelative features. Work here is poised to give the first insights into the autocorrelation structure of various raw and derived physiologic measures, multicollinearity of these features, the impact of data resolution on such aspects of data structures, and the overall computational considerations for derivation of optimal autocorrelative models. Understanding these fundamental aspects of physiologic data streams is required for robust platforms that forecast future states of physiological variables, building the foundation for more robust and complicated modelling methods.
Program
Alliance Grants
Recipient
Location
Manitoba, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data
As published
Application id
586244-2023
Name nom
Zeiler, Frederick FA
Department département
Surgery
Organization id
19
Institution établissement
University of Manitoba
Province en
Manitoba
Province fr
Manitoba
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2023
Competition year année de concours
2023
Award amount
25000
Program id
ALLRP
Program name en
Alliance Grants
Program name fr
Subventions Alliance
Group en
RESEARCH PARTNERSHIPS
Group fr
PARTENARIATS DE RECHERCHE
Committee code
96
Committee name en
RPP Internal Decision Cttee
Committee name fr
Cté de décision interne - PPR
Area of application code
1102
Area of application group en
Health, education and social services
Area of application group fr
Santé publique, éducation et services sociaux
Area of application en
Biomedical engineering
Area of application fr
Génie biomédical
Research subject code
6400
Research subject group en
Life sciences research related to human health and disease
Research subject group fr
Life sciences research related to human health and disease
Research subject en
Sciences de la vie liées à la santé publique et aux maladies humaines
Research subject fr
Sciences de la vie liées à la santé publique et aux maladies humaines
Application title
Autocorrelative and Multicollinearity Structures of Mammalian Cerebral Physiologic Data Streams for Future Application in Machine Learning Pipelines
Keyword
artificial intelligence; autocorrelation; cerebral physiology; computational time; machine learning; multicollinearity; statistical relationships; temporal resolution; time-series analytics; vector space
Application summary
Trajectory modelling of physiologic state (pre-emptively anticipating future states) is considered a 'holy grail' for various physiologic, data science and life/veterinary science fields. To facilitate this temporally resolved physiologic forecasting, modelling of the data is required which now often leverages machine learning/artificial intelligence driven processes. Typically, physiologic data, in current models, is inputted without accounting for inherent autocorrelative structures of parent signals or multicollinearity in vector space (the linear impact within interlinked data). Violating such priors of linearity leads to model errors and uncertainty in significance of forecasted values. Currently, our understanding of the time-series autocorrelative structures of cerebral physiology and the interconnected multicollinearity between them is unknown. Such work is further impaired by the impact of data resolution and computational lag time required for optimal physiological model creation. To improve physiological models and forecasting such knowledge gaps need to be fulfilled. Leveraging existing globally unique high-fidelity, high-frequency multi-modal cerebral physiologic data sets in humans from both the University of Manitoba (UM) and University of Helsinki (UH), this NSERC Alliance International Catalyst project aims to apply time-series methodologies (UM) and leverage collaborative expertise in their consideration for machine learning pipeline development (UM and UH), characterizing: the autocorrelative structure/multicollinearity features, the impact of data resolution, and computational considerations with derivation of models to account for such autocorrelative features. Work here is poised to give the first insights into the autocorrelation structure of various raw and derived physiologic measures, multicollinearity of these features, the impact of data resolution on such aspects of data structures, and the overall computational considerations for derivation of optimal autocorrelative models. Understanding these fundamental aspects of physiologic data streams is required for robust platforms that forecast future states of physiological variables, building the foundation for more robust and complicated modelling methods.