S16861Q001S
THE OBJECTIVE OF THIS EFFORT IS TO DEMONSTRATE THE USE OF DEEP TOPOLOGICAL MODELING TO CREATE AN EXPLAINABLE MACHINE LEARNING SOLUTION FOR ANALYZING MULTI-MODAL DATASETS OBTAINED FROM HUMAN NEUROVASCULAR UNIT-ON-A-CHIP (NVU-OOC) MODELS EXPOSED TO HYPOXIA CONDITIONS. THE NVU-OOC, A TYPE OF MICROPHYSIOLOGICAL SYSTEM (MPS), MODELS THE BASIC STRUCTURAL AND FUNCTIONAL UNITS, AS WELL AS THE PHYSIOLOGICAL MICROENVIRONMENT, OF THE HUMAN BRAIN. MPSS ARE COMPOSED OF ONE OR MULTIPLE CELL TYPES OF HUMAN OR ANIMAL ORIGIN THAT MIMIC THE BIOCHEMICAL, ELECTRICAL, AND MECHANICAL RESPONSES, AS WELL AS CELL BARRIER PROPERTIES, OBSERVED WITHIN ORGAN SYSTEMS. THESE IN VITRO SYSTEMS EXPEDITE AND ADVANCE SCIENTIFIC DISCOVERY WITH IMPROVED PREDICTIVE AND TRANSLATIONAL CAPABILITIES, WITHOUT THE USE OF ANIMAL MODELS OR HUMAN SUBJECTS. MPSS SERVE AS ORGAN SURROGATES TO TEST THE EFFECTS OF VARIOUS STRESSORS/STIMULI (E.G., HEAT STRESS, HYPOXIA, DRUGS). RESULTS FROM THESE SYSTEMS CAN BE USED TO DEVELOP INTERVENTIONAL AND COUNTERMEASURE STRATEGIES, AND TAILOR AND SCOPE ADDITIONAL STUDIES USING ANIMAL MODELS OR HUMAN SUBJECTS IF NEEDED. MPSS CAN BE CUSTOMIZED TO REFLECT INDIVIDUAL PHYSIOLOGY, SUCH AS THAT OF WARFIGHTERS, BY CREATING ORGANOIDS DERIVED FROM PLURIPOTENT STEM CELL (IPSC) TECHNOLOGY âÂ'¬Â€Œ A PATH TOWARDS CREATING HUMAN DIGITAL TWINS. NVUOOCS ARE SUBJECTED TO DYNAMIC OXYGEN EXPOSURES THAT SIMULATE BRAIN-SPECIFIC HYPOXIA, HYPEROXIA, OR NORMOXIA CONDITIONS FOR DEFINED TIME PERIODS. VARIOUS DATA TYPES, INCLUDING CELLULAR HEALTH METRICS, FUNCTIONAL ASSAYS, MULTI-OMICS, AND OTHER METADATA, HAVE BEEN COLLECTED FROM THESE EXPERIMENTS. OUR DEEP TOPOLOGICAL MODELING CAPABILITIES, WHICH EXTEND DIMENSIONALITY REDUCTION TECHNIQUES AND METHODS, WILL BE USED TO FORMULATE A ROBUST UNDERSTANDING OF THE MULTIVARIATE SIGNATURES WITHIN THE NVUOOC DATA. MINED XAI HAS DEVELOPED DEEP TOPOLOGICAL MODELING CAPABILITIES BY EXTENDING DIMENSIONALITY REDUCTION TECHNIQUES AND METHODS, USING OPEN ARCHITECTURE DESIGN PRINCIPLES TO FORMULATE A ROBUST UNDERSTANDING OF COMPLEX SIGNALS, WHICH CAN BE GENERALIZED ACROSS DOMAINS. WE HAVE FURTHER REVOLUTIONIZED AND MATURED DEEP TOPOLOGICAL MODELING TO PROVIDE A FULLY EXPLAINABLE EXPLOITATION TOOL WHEREBY MODELS THROUGHOUT THE ARCHITECTURE CAN BE ANALYZED EITHER IN COMPONENT OR IN AGGREGATE. OUR XAI MODELS CAPTURE THE COMPLEX MULTIVARIATE SIGNATURES PRESENT IN BIOLOGICAL SYSTEMS AND PROVIDE A COMPREHENSIVE VIEW OF THE UNDERLYING DATA, ENABLING A DEEPER UNDERSTANDING OF THE INTRICATE BIOLOGICAL PROCESSES OCCURRING WITHIN THE NVU-OOC MODELS. NVU-OOC GENERATE MULTI-MODAL DATA FROM VARIOUS BIOLOGICAL, PHYSIOLOGICAL, AND MECHANICAL MEASUREMENTS CONDUCTED VIA A SET OF TIME SERIES EXPERIMENTS. ANALYZING MULTIVARIATE DATA OBTAINED FROM NVU-OOC WITH EXPLAINABILITY CAN BE CHALLENGING TO INTEGRATE AND INTERPRET. THE APPLICATION OF XAI, SPECIFICALLY THROUGH OUR HIERARCHICAL DEEP TOPOLOGICAL DATA MODELING, OFFERS A POWERFUL TOOL FOR ANALYZING MULTI-MODAL DATA FROM NVUOOC. BY PROVIDING SCALABILITY, EXPLAINABILITY, AND COMPREHENSIVE MULTIVARIATE ANALYSIS, THIS APPROACH WILL ENHANCE THE UNDERSTANDING OF COMPLEX BIOLOGICAL NETWORKS, ULTIMATELY LEADING TO MORE INFORMED AND RELIABLE SCIENTIFIC CONCLUSIONS.
Mined XAI LLC
Delivery Order FA865019D6109-FA238423FB001
$258.8k 8/22/24 S16861X001S
THIS PROSPECTIVE STUDY WILL ENROLL 60 PARTICIPANTS AT EITHER WPAFB/WSU OR WHASC (WILFORD HALL AMBULATORY SURGICAL CENTER). PARTICIPANTS MUST HAVE AN AFFILIATION WITH EITHER DOD OR WSU, BE BETWEEN 18 AND 42 YEARS OLD, MEET STANDARD ELIGIBILITY REQUIREMENTS FOR BOTH MRI (E.G., IMPLANTED OR NON-REMOVABLE METAL, PREGNANCY, HISTORY OF ANEURYSM, CLAUSTROPHOBIA, ETC.) AND NEUROSCIENTIFIC (E.G., DIAGNOSIS OF A NEUROLOGICAL AND/OR PSYCHOLOGICAL CONDITION, USE OF PSYCHOTROPIC MEDICATIONS, SLEEP DISORDERS, MIGRAINE OR FREQUENT HEADACHES, FREQUENT EYE STRAIN, EPILEPSY, LEARNING DISABILITY, ETC.) STUDIES. ADDITIONALLY, BECAUSE TVNS WILL BE EMPLOYED, POTENTIAL PARTICIPANTS WILL BE EXCLUDED IF THEY HAVE HAD VAGOTOMY OR ABNORMAL ANATOMY AT THE SITE OF THE TVNS STIMULATION (NECK), SUCH AS SCARRING, WOUNDS, SWELLING, OR INFECTION OF THE SKIN AT THE STIMULATION SITE. THE STUDY WILL EMPLOY A REPEATED-MEASURES DESIGN, IN WHICH 2 LEVELS EACH OF 2 VARIABLES (TREATMENT GROUP: TVNS VS. SHAM, TRAINING LEVEL: PRE- VS. POST-TRAINING) WILL BE COMPARED. PER A POWER ANALYSIS COMPLETED BY AFRL/WSU PERSONNEL, 60 PARTICIPANTS SHOULD BE SUFFICIENT TO DETECT AN EFFECT BASED UPON THE EXPECTED EFFECT SIZE IN 42 PARTICIPANTS WHO COMPLETE THE ENTIRETY OF PARTICIPATION IN THE STUDY (ACCOUNTING FOR ATTRITION, ETC.). THE TASK DURING WHICH TRAINING AND TVNS EFFECTS WILL BE MEASURED WILL BE A VERSION OF THE SAR TASK USED PREVIOUSLY THAT HAS BEEN MODIFIED FOR USE IN THE FMRI ENVIRONMENT.
KBR Wyle Services, LLC
Delivery Order FA865019D6109-FA238423FB001
$564.2k 6/1/24