Research & Publications

Physics, Signals, Disease and AI

My work sits at the intersection of imaging physics, signal processing and machine learning, with a focus on making quantitative brain imaging reliable. At the Imaging Genetics Center at USC, developing diffusion MRI pipelines for ADNI showed me how differences in scanner hardware, acquisition protocols and reconstruction affect not only image contrast but also the methods we use to analyze it, which led me to help develop a style-encoding GAN for multisite MRI harmonization that preserves anatomy. At Washington University, I investigated how the scanner's point-spread function causes signal to spill across anatomical boundaries in preclinical PET, and developed a geometry-adaptive sampling method to reduce partial-volume bias. At UCSF's Memory and Aging Center, I led MRI-based PET processing and designed the quality-control system for the multisite LEADS study, where the measures I generated informed clinical decisions and the evaluation of patients receiving anti-amyloid therapy. Across these projects, I have learned that a measurement is only as reliable as our understanding of how it was acquired, processed and validated. That understanding guides my current work building clinical AI tools at Whissle.ai and my independent research on ultrashort echo time MRI of myelin, and reflects my long-term goal to engineer robust and mathematically principled imaging systems for research and precision diagnostics.

Projects

Research Projects

Tools and methods I've built to make quantitative brain imaging more reliable, from PET quantification to multisite MRI harmonization.

Inflated skeleton of the mouse cerebellum colored by PET uptake, raw and noise plus smoothing images, with the mean bias of the inflation over five phantom brains: minus 0.2 and minus 12.6 percent Inflated skeleton of the mouse cerebellum colored by PET uptake, raw and noise plus smoothing images, with the mean bias of the inflation over five phantom brains: minus 0.2 and minus 12.6 percent

3D Skeleton-Based PET Quantification

In development

Small mouse-brain regions read too low in PET because of the scanner's limited resolution and small registration errors. I sample each region along its 3D skeleton, inflated with spheres sized to the local thickness, so the measurement stays away from the region's edge and needs no scanner model. Validated on digital phantoms with known activity.

Preclinical PETPartial-volume effect3D geometryPython

For access or collaboration, contact Dr. Kooresh Shoghi (Shoghi Lab, Washington University School of Medicine) or Piyush Maiti.

Style-Transfer GAN for Multi-site MRI Harmonization

MICCAI 2021

MRI scans pooled from many sites differ in contrast and intensity because of the scanner, not the brain. We treat harmonization as style transfer: a GAN based on StarGAN v2 takes the scanner "style" from a single reference scan and applies it to any brain while keeping the anatomy, with no site labels or traveling subjects. Trained on 718 T1-weighted scans from five cohorts (UK Biobank, ADNI, PPMI, ABCD, ICBM); it raised segmentation Dice on an unseen dataset and worked on unseen scanners. Top row: original scans; bottom row: harmonized.

MRI harmonizationGANsStyle transferPyTorch
GitHub ↗ Paper ↗

Publications & Abstracts

13
Publications
18
Abstracts
298
Citations
6
H-Index
Google Scholar Profile

Distinct Tau PET Dynamics in Early- Versus Late-Age-of-Onset Alzheimer's Disease

Journal

Chiotis K., Blazhenets G., Soleimani-Meigooni D. N., …, Maiti P., …, et al.

Science Translational Medicine, 18(874)•2026•eaea2816
DOI
Tau PETEarly-Onset ADLEADS

Hemispheric Asymmetry of Neuroimaging Biomarkers in Early Onset Alzheimer's Disease: Clinical and Cognitive Outcomes

Journal

Ziontz J., Blazhenets G., Visani A., Chiotis K., Maiti P., Lin Z., et al.

Annals of Neurology•2026
Full TextDOI
Early-Onset ADHemispheric AsymmetryBiomarkers

Identification of Patients Receiving Amyloid-Targeting Therapies in Observational Studies Using Amyloid PET Trajectories: Insights from LEADS

Journal

La Joie R., Blazhenets G., Maiti P., Chiotis K., Eloyan A., Kirby K., et al.

Alzheimer's & Dementia: DADM, 18(2)•2026•e70408
Full TextDOI
Amyloid PETAnti-Amyloid TherapyLEADS

Tau Topography Subtypes Account for Clinical Heterogeneity and Longitudinal Trajectories in Early-Onset Alzheimer's Disease

Journal

Lin M., Chiotis K., Maiti P., Zhang J., Blazhenets G., et al.

Brain Communications, 8(3)•2026•fcag176
Full TextDOI
Tau PETEarly-Onset ADSubtypes

Tablet-Based Cognitive Assessment Tool (TabCAT) in a Multisite Study of Early-Onset Alzheimer's Disease

Journal

Njamnshi W. Y., Chiotis K., Tsoy E., Blazhenets G., Maiti P., Zhang J., et al.

Alzheimer's & Dementia, 21(12)•2025•e71012
Full TextDOI
Cognitive AssessmentTabCATEarly-Onset AD

Exploring Centiloid Robustness: Impact of Sample Size and Image Resolution on Centiloid Conversion Accuracy

Journal

Zhang J., Soleimani-Meigooni D. N., Koeppe R., Ackley S., Schonhaut D. R., Lin Z. E., Maiti P., et al.

Journal of Nuclear Medicine, 66(12)•2025•pp. 1976–1984
Full TextDOI
PETCentiloidQuantification

Amyloid PET in Sporadic Early- Versus Late-Onset Alzheimer's Disease: Comparison of the LEADS and ADNI Cohorts

Journal

Lagarde J., Maiti P., Schonhaut D. R., Blazhenets G., Zhang J., Eloyan A., Thangarajah M., et al.

Annals of Neurology, 98(2)•2025•pp. 236–248
Full TextDOI
Alzheimer's DiseasePETLEADSADNI

Evaluation of Copanlisib in Combination with Eribulin in Triple-Negative Breast Cancer Patient-Derived Xenograft Models

Journal

Guo Z., Luo J., Mashl R. J., Hoog J., Maiti P., Fettig N., Davies S. R., et al.

Cancer Research Communications, 4(6)•2024•pp. 1430–1440
Full TextDOI
Cancer ResearchPreclinical ImagingPDX Models

Style Transfer Generative Adversarial Networks to Harmonize Multisite MRI to a Single Reference Image to Avoid Overcorrection

Journal

Liu M., Zhu A. H., Maiti P., Thomopoulos S. I., Gadewar S., Chai Y., Kim H., Jahanshad N.

Human Brain Mapping, 44(14)•2023•pp. 4875–4892
Full TextDOI
Deep LearningMRI HarmonizationGANs

Application of Ultrasonic Sensors for the Development of Anti-Collision Device

Journal

Maiti P., Bhusnur S.

Int. J. Electrical and Electronic Eng. & Telecomm.•2016•Vol. 5, No. 1, pp. 20–27
SensorsEngineering

Style Transfer Using Generative Adversarial Networks for Multi-Site MRI Harmonization

Conference

Liu M., Maiti P., Thomopoulos S., Zhu A., Chai Y., Kim H., Jahanshad N.

MICCAI 2021•2021•pp. 313–322, Springer
DOI
Deep LearningMRI HarmonizationGANs

Diffusion MRI Metrics and Their Relation to Dementia Severity: Effects of Harmonization Approaches

Conference

Thomopoulos S. I., Nir T. M., Villalon-Reina J. E., Zavaliangos-Petropulu A., Maiti P., Zheng H., et al.

SIPAIM 2021 (Proc. SPIE)•2021•Brazil
DOI
Diffusion MRIHarmonizationDementia

Region-Specific Automatic Quality Assurance for MRI-Derived Cortical Segmentations

Conference

Gadewar S., Zhu A. H., Thomopoulos S. I., Li Z., Ba Gari I., Maiti P., Thompson P. M., Jahanshad N.

IEEE ISBI 2021•2021•pp. 1288–1291
DOI
Quality ControlMRISegmentation

Discordance Between Amyloid PET Visual Read and Quantitation: False Positives or Early Pathology Detection?

Abstract

Heston M. B., Maiti P., Blazhenets G., Soleimani-Meigooni D. N., Lin Z., et al.

Human Amyloid Imaging (HAI)•2026
Amyloid PETVisual ReadQuantification

Amyloid-PET and White Matter Hyperintensities but Not Military Service Duration nor Combat Exposure Are Associated with Neurodegeneration and Cognitive Impairment in Older Veterans

Abstract

Rocha S., Pasternak S., Lesman-Segev O., Schwartz D., Blazhenets G., Maiti P., et al.

Human Amyloid Imaging (HAI)•2026
Amyloid PETVeterans HealthWhite Matter

Tau Topography Subtypes in Early-Onset Alzheimer's Disease: Explaining Clinical Heterogeneity and Propagation Patterns

Abstract

Lin M., Maiti P., Zhang J., Blazhenets G., Rocha S., Shankar R., et al.

AAIC 2025•2025•Alzheimer's & Dementia 21(S2): e105737
DOI
Tau PETEarly-Onset ADSubtypes

Target Engagement and Cognitive Outcomes of Anti-Amyloid Immunotherapies in Early-Onset Alzheimer's Disease: First Report from the LEADS Study

Abstract

Chiotis K., Blazhenets G., Eloyan A., Maiti P., Zhang J., Touroutoglou A., et al.

AAIC 2025•2025•Alzheimer's & Dementia 21(S2): e104661
DOI
Anti-Amyloid TherapyEarly-Onset ADLEADS

Characterizing Patients with Early-Onset Alzheimer's Disease and Amyloid-Positive, Tau-Negative (A+T−) PET Profiles: Data from LEADS

Abstract

Raya M. A., Soleimani-Meigooni D. N., Blazhenets G., Chiotis K., Lagarde J., Vrillon A., Maiti P., et al.

AAIC 2025•2025•Alzheimer's & Dementia 21(S2): e103476
DOI
Amyloid PETTau PETEarly-Onset AD

Clinical and Cognitive Manifestations of Hemispheric Asymmetry in Neuroimaging Biomarkers in Early-Onset Alzheimer's Disease

Abstract

Ziontz J., Maiti P., Zhang J., Chiotis K., Blazhenets G., Rocha S., Shankar R., et al.

AAIC 2025•2025•Alzheimer's & Dementia 21(S3): e101192
DOI
Early-Onset ADHemispheric AsymmetryBiomarkers

Tau PET Load in Early- and Late-Onset Alzheimer's Disease: A Cross-Sectional and Longitudinal Comparison of the LEADS and ADNI Cohorts

Abstract

Chiotis K., Blazhenets G., Schonhaut D. R., Lagarde J., Soleimani-Meigooni D. N., Maiti P., Zhang J., et al.

AAIC 2025•2025•Alzheimer's & Dementia 21(S2): e104867
DOI
Tau PETLEADSADNI

Tablet-Based Cognitive Assessment in a Multisite Study of Early-Onset Alzheimer's Disease: Association with Pen-and-Paper Tests and PET Measures of Amyloid and Tau

Abstract

Njamnshi W. Y., Chiotis K., Tsoy E., Possin K. L., Hammers D. B., Eloyan A., Kirby K., Maiti P., et al.

AAIC 2025•2025•Alzheimer's & Dementia 21(S3): e104654
DOI
Cognitive AssessmentTabCATEarly-Onset AD

Characterization of the Heterogeneity of Amyloid-PET-Negative Patients with a Clinical Diagnosis of Sporadic Early-Onset AD: An FDG-PET Study in the LEADS Cohort

Abstract

Lagarde J., Schonhaut D. R., Maiti P., Zhang J., Soleimani-Meigooni D. N., Zeltzer E., et al.

AAIC 2024•2024•Alzheimer's & Dementia 20(S2): e092804
DOI
FDG-PETEarly-Onset ADLEADS

Amyloid-PET in Patients with a Clinical Diagnosis of Sporadic Early- Versus Late-Onset AD: Comparison of the LEADS and ADNI Cohorts

Abstract

Lagarde J., Maiti P., Schonhaut D. R., Zhang J., Soleimani-Meigooni D. N., Zeltzer E., et al.

AAIC 2024•2024•Alzheimer's & Dementia 20(S9): e094101
DOI
Amyloid PETLEADSADNI

A Clock Model of Regional Tau-PET Accumulation and Cognitive Decline in Sporadic Early-Onset Alzheimer's Disease

Abstract

Schonhaut D. R., Insel P. S., Soleimani-Meigooni D. N., Zeltzer E., Windon C., Maiti P., Yballa C., et al.

AAIC 2024•2024•Alzheimer's & Dementia 20(S9): e093935
DOI
Tau PETEarly-Onset ADCognitive Decline

Assessing the Performance of Software Platforms to Segment Mouse Brains Using a Digital Phantom

Abstract

Maiti P., Li S., Dutta K., Whitehead T., Shoghi K.

SNMMI 2023•2023•J Nucl Med 64(S1): P1362
Abstract
Preclinical ImagingSegmentationQuality Control

Imaging Neuroinflammation in an Animal Model of Diet-Induced Obesity

Abstract

Shyam C., Whitehead T., Zhou D., Qiu L., Maiti P., Tu Z., Kravitz A., Hershey T., Shoghi K.

SNMMI 2023•2023•J Nucl Med 64(S1): P1356
Abstract
PETNeuroinflammationObesity

Effects of Dementia and MCI on Diffusion Tensor Metrics Using the Updated ADNI3 DTI Preprocessing Pipeline

Abstract

Thomopoulos S. I., Nir T. M., Villalon-Reina J. E., Zavaliangos-Petropulu A., Maiti P., et al.

AAIC 2022•2022•Alzheimer's & Dementia 18(S6): e066333
DOI
Diffusion MRIADNI3Dementia

Cortical Microstructural Associations with CSF Amyloid and Tau

Abstract

Nir T. M., Villalon-Reina J. E., Haddad E., Zheng H., Thomopoulos S. I., Maiti P., et al.

AAIC 2021•2021•Alzheimer's & Dementia 17(S4): e055820
DOI
Diffusion MRICSF BiomarkersAmyloid

Multi-Site MRI Harmonization Using Unified Generative Adversarial Networks

Abstract

Liu M., Maiti P., Thomopoulos S., Kadakova N., Kim H., Jahanshad N.

OHBM 2021•2021•Seoul, Korea
MRIHarmonizationGANs

A 3D U-Net for Gibbs Artifact Removal from Quantitative Susceptibility Maps

Abstract

Ba Gari I., Gadewar S. P., Wei X., Maiti P., Boyd J., Jahanshad N.

ISMRM 2021•2021
Deep LearningMRIArtifact Removal

Hippocampal Subfield Microstructure Abnormalities Mediate Associations Between Tau Burden and Memory Performance

Abstract

Nir T. M., Salminen L., Villalon-Reina J. E., Tubi M. A., Thomopoulos S. I., Maiti P., Braskie M. N., Thompson P. M., Jahanshad N.

AAIC 2020•2020•Alzheimer's & Dementia 16(S4): e039622
DOI
Diffusion MRIHippocampusTau