Research

Reading the brain — and reading it well

My lab develops signal-processing, modeling, and machine-learning methods for human (and canine) neuroimaging, and applies them to questions in cognitive, social, clinical, and comparative neuroscience. Below is the work, organized into the methods we build and the domains in which we apply them.

Part 01

Methodological innovations

The methods side of the lab focuses on extracting, modeling, and interpreting signals from functional brain imaging data—particularly fMRI—at the limits of current technology. We build network models of brain function, design AI architectures that work on noisy, multi-site neuroimaging data, and push toward sub-millimeter, layer-resolved imaging at 7 Tesla.

  • Functional Magnetic Resonance Imaging (fMRI)
  • Signal & image processing
  • Network modeling: autoregressive, Bayesian, wavelet, state-space
  • Big data analytics in neuroimaging
  • Pattern recognition & deep learning
  • Multimodal imaging (EEG / fMRI fusion)
  • Ultra-high-field 7T MRI
  • Sub-mm layer fMRI
  • Multinuclear MR spectroscopy
  • Real-time fMRI neurofeedback
  • Hemodynamic response function modeling
  • Granger causality & effective connectivity
Part 02

Applied neuroscience

On the application side, our methods are deployed across an unusually broad set of domains—from the basic neuroscience of touch and vision, to the social neuroscience of trust and altruistic punishment, to the clinical neuroscience of PTSD, schizophrenia, autism, and Alzheimer’s. We also do unconventional comparative work in awake, unrestrained dogs.

  • Sensory, motor & cognitive neuroscience
  • Social & affective neuroscience
  • Consciousness
  • Meditation & complementary medicine
  • Autism Spectrum Disorder
  • Alzheimer’s & mild cognitive impairment
  • Post-traumatic stress disorder (PTSD)
  • Mild traumatic brain injury (mTBI)
  • Schizophrenia & psychosis
  • Parkinson’s disease & depression
  • Neuroeconomics & neuromarketing
  • fMRI in awake working dogs

Selected external funding

$50M+ as PI / Co-PI
  • DHS

    DCSITE — Detection Canine Sciences, Innovation, Technology and Education

    2022 – 2026 · Co-PI (PI: Bartol, AU)

    $49.9MCo-PI
  • NIH R01

    Cross-modal correspondences between visual and auditory features

    2015 – 2020 · Multiple-PI

    $2.26MMPI
  • DARPA STTR

    Functional imaging for developing outstanding service dogs (Phase 2)

    2014 – 2017 · Principal Investigator

    $996KPI
  • NSF

    Applications in parallel and distributed computing

    2017 – 2020 · Co-PI (PI: Baskiyar, AU)

    $458KCo-PI
  • Varian Med.

    Pilot trial of frameless virtual cone stereotactic radiosurgical thalamotomy for intractable tremor

    2017 – 2019 · Co-PI (PI: Bredel, UAB)

    $311KCo-PI
  • AFOSR

    Neural signatures of trust during human–automation interactions

    2013 – 2016 · Multiple-PI

    $150KMPI
  • Army Research Office

    MRI brain imaging of post-concussion syndrome

    2012 – 2013 · Co-PI (PIs: Denney, AU; Dretsch, Walter Reed)

    $119KCo-PI
  • DoD

    Examining oxytocin as a causal mechanism for long-term bonding between humans and autonomy

    2021 – 2023 · Co-PI (PI: Krueger, GMU)

    $88KCo-PI
  • PAIR (Auburn)

    Establishment of the Center for Neuroscience

    2018 – 2021 · Co-PI (PI: Suppriramaniam, AU)

    $637KCo-PI

Patent

USA Patent · No. 11013426 · Issued 5 May 2021

System and method of functional MRI of the neural system in conscious, unrestrained dogs

G. Deshpande, P. Waggoner, V. Vodyanoy, H. Jia, O. Pustovyy, T. Denney, E. Morrison, R. Beyers