Regional perfusion maps vulnerability, not yet symptom trajectories, in prodromal dementia

Question: How do arterial spin labeling and other perfusion MRI methods relate regional cerebral blood flow changes to neuropsychiatric symptom trajectories in prodromal dementia cohorts?

Answer

Across the retrieved literature, the honest answer is that this relationship has not been demonstrated. The two evidence streams run in parallel rather than intersecting: longitudinal ASL work in the AD spectrum shows that regional hypoperfusion tracks tau pathology in posterior cortex, independent of amyloid [7], while trajectory-class analyses of neuropsychiatric symptoms (NPS) in prodromal cohorts are built on structural, diffusion, and plasma markers with no perfusion term at all [14], [13], [11]. Where CBF is tested directly against behaviour, associations are weak or absent while flow-derived or non-CBF vascular metrics carry the signal [1], [8]. Taken together, these papers position perfusion as a spatially organised map of tissue vulnerability [4], [5] from which symptom-level inferences are currently being made by assumption rather than by measurement.

Longitudinal ASL coupling has been demonstrated for tau, not for behaviour

The strongest longitudinal perfusion result in a prodromal-to-dementia cohort ties CBF to proteinopathy rather than to symptoms: in 179 ADNI3 participants, tau PET was negatively associated with ASL-derived CBF both cross-sectionally and longitudinally in posterior brain, independent of amyloid PET [7]. That same study found white matter hyperintensity associations with amyloid and tau attenuated after mutual adjustment, which the authors interpret as shared disease-related variance rather than independent effects [7]. This means the one well-characterised longitudinal CBF axis available for prodromal cohorts is a pathology axis, and any NPS trajectory claim must be justified against it.

NPS trajectory classes in prodromal cohorts are currently defined without perfusion

Growth mixture modelling of NPI data in 785 ADNI participants identified four NPS trajectory classes that predicted cognitive decline over time, yet these classes were unrelated to longitudinal white matter microstructure of the cerebellar peduncles in the 60 individuals with adequate imaging [14]. Parallel work on the apathy domain anchors it to fluid and structural markers instead: MBI-apathy showed higher plasma p-tau181 at baseline and over two and three years versus no-NPS [13], and MBI-apathy was interrogated against AD meta-regions-of-interest cortical thickness and grey matter volume composites [11]. Neither body of work includes a perfusion measurement, so the CBF–NPS trajectory link is untested rather than negative.

Direct tests of CBF against cognition are negative or conditional, and flow-derived metrics dissociate

In 36 older adults spanning cognitively unimpaired, MCI, and AD, CBF correlated with a SNAP-derived distal cerebral arterial flow index but showed no significant correlation with RBANS total score, while the arterial flow index itself did [1]. Cerebrovascular reactivity measured with ASL behaved similarly weakly: correlations between TE-independent noise and ASL-CVR were small and negative (approximately −0.11 to −0.14) with no significant differences between healthy controls, MCI, and AD, although group differences did emerge within specific functional networks [2]. Even in healthy ageing, cardiorespiratory fitness–CBF associations in posterior cingulate, anterior cingulate, precuneus, and hippocampus disappeared after adding body mass index, and were absent in a second cohort [18]. Perfusion–behaviour effect sizes in these cohorts are therefore fragile to third variables and to which flow parameter is chosen.

Where regional perfusion does earn its keep, it is as a normative deviation and vulnerability map

A normative model built on over 12,000 ASL scans quantified individual deviations from expected CBF across neurodegenerative and psychiatric conditions and used longitudinal data to establish typical and atypical perfusion trajectories, arguing explicitly for prognostic value in detecting disease progression [5]. Perfusion topography also predicts where degeneration lands: cortical atrophy in late-onset AD, TDP-43C, and dementia with Lewy bodies was most pronounced in regions with lower perfusion, supporting a transdiagnostic vulnerability interpretation [4]. At population scale, ASL in UK Biobank showed significant associations with health information including depression and was more sensitive to many factors than other imaging modalities [3]. This is the correct inferential frame for NPS: regional CBF may index which networks are exposed, not how severe a symptom is today.

The NPS-adjacent vascular signals that do replicate are not CBF

Reduced BOLD-CSF coupling — a glymphatic rather than perfusion metric — was associated with hallucination severity, cognitive fluctuations, poorer sleep quality, and impaired attentional working memory in Parkinson's disease and dementia with Lewy bodies, but not with motor burden [8]. Choroid plexus blood flow measured with pseudo-continuous ASL declined over two years in 90 Parkinson's patients, and that decline was associated with worsening cognitive performance [10]. Progressive white matter hyperintensity and perivascular space volumes were associated with cortical atrophy and cognitive decline over ten years [9]. Affective NPS are also more frequent in Lewy body disorders than in AD or vascular cognitive impairment regardless of the metric used, indicating that symptom expression is conditioned on which substrate is affected [15].

Conflicts and gaps

The causal direction of hypoperfusion is unresolved: a lumped-parameter hydrodynamic model argues that in MCI capillary transmural pressure is normal but falls as AD progresses, and proposes that the AD brain renders itself ischaemic, partially mitigating blood-brain barrier disruption — hypoperfusion as compensation, not injury [6]. That directly contradicts reading low CBF as damage severity. Methodologically, transit-time and flow metrics dissociate: after acetazolamide, arterial transit time and macrovascular transit time both fell, but changes in transit time were not associated with cerebrovascular reactivity whereas labelled blood volume change was [16]. Cross-cohort pooling requires harmonisation, which reduced cohort differences in ASL features and improved ASL-only brain-age error from 11.1 to under 8.8 years [12]. NPS measurement itself is unstable — agreement on depression presence across metrics was 6% and on anxiety 7% in NACC [15] — and mechanistic work on apathy is proceeding through prefrontal GABAergic gain and prior precision rather than through perfusion [17]. Sample sizes in the perfusion-behaviour papers above are small (36; 39; 179), and none report NPS trajectories as an endpoint.

What would change the answer

A longitudinal prodromal cohort with baseline and repeat multi-delay pCASL, repeated NPI-Q or MBI-C, and latent class/growth mixture trajectory modelling would settle the core question, ideally with tau PET or plasma p-tau217 as a mediator and CBF expressed as normative z-scores rather than absolute units [5]. Arterial transit time and cerebrovascular reactivity must enter as separate regressors rather than being folded into CBF [16], and harmonisation across sites is a precondition for pooling [12]. A domain-specific test is now feasible: if perfusion marks vulnerability topography [4], apathy should localise to anterior cingulate and ventromedial prefrontal perfusion while hallucinations localise to posterior and thalamic territory, a prediction testable in ADNI or UK Biobank repeat-ASL data [3]. Finally, competing vascular markers — BOLD-CSF coupling [8], choroid plexus perfusion [10] — should be modelled jointly with cortical CBF, because on current evidence it is these, not regional CBF, that carry the symptom signal.

References

[1] Ma, X., Koppelmans, V., Akcicek, H., Akcicek, E. Y., Shen, J., Chen, L., Balu, N., Yuan, C., King, J. B.. "SNAP MRI Reveals Association Between Distal Cerebral Arterial Flow and Cognitive Function in an Aging Population". bioRxiv. 2026-03-25T00:00:00.000Z. URL: https://doi.org/10.64898/2026.03.21.713392

[2] Guidi, M., Giulietti, G., Singh, T., Mancini, M., DiNuzzo, M., Bonarota, S., Caruso, G., Di Domenico, C., Esposito, F., Serra, L., Caltagirone, C., Carlesimo, G. A., Giove, F.. "Non-neuronal signal fluctuations in Alzheimer's disease and in mild cognitive impairment". bioRxiv. 2026-07-24T00:00:00.000Z. URL: https://doi.org/10.1101/2025.10.01.679773

[3] Okell, T. W., Xu, X., Craig, M., Alfaro-Almagro, F., Thomas, D. L., De Vita, E., Garratt, S., Nichols, T. E., Guenther, M., Matthews, P. M., Miller, K. L., Smith, S. M., Chappell, M. A.. "Brain Perfusion Imaging of a Large Population: Arterial Spin Labelling MRI in UK Biobank". medRxiv. 2025-08-08T00:00:00.000Z. URL: https://doi.org/10.1101/2025.07.23.25332042

[4] Farahani, A., Liu, Z.-Q., Ceballos, E. G., Hansen, J. Y., Wennberg, K., Zeighami, Y., Dadar, M., Gauthier, C. J., Dagher, A., Misic, B.. "Cerebral blood perfusion across biological systems and the human lifespan". bioRxiv. 2025-02-06T00:00:00.000Z. URL: https://doi.org/10.1101/2025.02.05.636674

[5] Xinglin Zeng, Yiran Li, Lin Hua, Ruoxi Lu, Lucas Lemos Franco, Peter Kochunov, Shuo Chen, John A Detre, Ze Wang. "Normative Cerebral Perfusion Across the Lifespan". arxiv. 2025-02-13T00:00:00.000Z. URL: https://arxiv.org/abs/2502.0807

[6] Bateman, G. A., Bateman, A. R.. "Brain Ischemia in Alzheimer's disease counteracts the Disruption of the Blood Brain Barrier: A Hypothesis Investigated with a Lumped Parameter Model.". medRxiv. 2024-12-20T00:00:00.000Z. URL: https://doi.org/10.1101/2024.12.18.24319268

[7] Lin, K., Sachdev, P., Jiang, J., Alzheimer's Disease Neuroimaging Initiative,. "The Associations of Cerebral Blood Flow and White Matter Hyperintensities with Tau and Amyloid-beta Across the Alzheimer's Disease Spectrum". medRxiv. 2026-05-27T00:00:00.000Z. URL: https://doi.org/10.64898/2026.05.25.26354067

[8] Ignatavicius, A., Konuri, A., Churchill, L., Anderson, J., Halliday, G., Lewis, S. J., Matar, E.. "Impaired cerebrovascular-cerebrospinal fluid coupling is associated with non-motor features of Lewy body disease". medRxiv. 2026-09-09T00:00:00.000Z. URL: https://doi.org/10.64898/2026.05.13.26353090

[9] Chen, Z., Raeesi, S., Metz, A., Moqadam, R., Chadwick, K. M., Zeighami, Y., Dadar, M.. "Contribution of cerebrovascular lesions to neurodegeneration and cognitive decline". medRxiv. 2026-10-06T00:00:00.000Z. URL: https://doi.org/10.64898/2026.10.04.26364698

[10] Li, H., Jia, J., Wang, J., Hu, X., Shao, X., Liu, K., Chen, J., Chen, Z., Jin, L., Wang, H.. "Complementary choroid plexus and locus coeruleus dysfunction in Parkinson's disease progression". medRxiv. 2026-07-06T00:00:00.000Z. URL: https://doi.org/10.64898/2026.07.02.26357179

[11] Vellone, D., Guan, D. X., Goodarzi, Z., Forkert, N. D., Smith, E. E., Ismail, Z.. "Apathy in Mild Behavioural Impairment: Associations with Cortical Thickness and Grey Matter Volume". medRxiv. 2026-02-27T00:00:00.000Z. URL: https://doi.org/10.64898/2026.02.25.26347107

[12] Dijsselhof, M. B. J., Moore, C., Amiri, S., Tee, M., Hilal, S., Chen, C., van den Born, B.-J. H., Nordhoy, W., Andreassen, O. A., Westlye, L. T., Chaturvedi, N., Hughes, A. D., Cash, D. M., Schott, J. M., Sudre, C. H., Barkhof, F., Kuijer, J. P. A., Biondo, F., Cole, J. H., Mutsaerts, H. J. M. M., Petr, J.. "Multi-cohort, multi-sequence harmonisation for cerebrovascular brain age". bioRxiv. 2025-04-29T00:00:00.000Z. URL: https://doi.org/10.1101/2025.04.22.649344

[13] Vellone, D., Leon, R., Goodarzi, Z., Forkert, N. D., Smith, E. E., Ismail, Z.. "Mild Behavioural Impairment-Apathy and Alzheimer's Disease Plasma Phosphorylated Tau Biomarker Levels". medRxiv. 2026-02-27T00:00:00.000Z. URL: https://doi.org/10.64898/2026.02.25.26347102

[14] Luisa Schmidt, Juan Quinones, Kristina Meyer, Andrea Hildebrandt. "Neuropsychiatric symptom trajectories and their associations with cognitive impairment and white matter microstructure of the cerebellar peduncles in Alzheimer’s disease". psyarxiv. 2025-10-27T21:22:57.552Z. URL: https://osf.io/a2wfb_v1

[15] Fisher, D. W., Mehta, R., Morrow, C. B., Kerr, K. F., Jayadev, S., Domoto-Reilly, K., Schrift, M. J., Darvas, M.. "Affective Neuropsychiatric Symptom Metrics in the National Alzheimers Coordinating Center Dataset". medRxiv. 2025-10-24T00:00:00.000Z. URL: https://doi.org/10.1101/2025.10.23.25338479

[16] Deckers, Q., Uniken Venema, S. M., Braun, K., van der Zwan, B., Deckers, P. T., Siero, J. C. W., Bhogal, A.. "Quantification of arterial hemodynamics in steno-occlusive disease using time-resolved MRI-based angiography". medRxiv. 2026-04-30T00:00:00.000Z. URL: https://doi.org/10.64898/2026.04.22.26350771

[17] Williams, R. S., Naessens, M., Jafarian, A., Lanskey, J., Hezemans, F. H., Friston, K., Hughes, L. E., Rowe, J.. "Apathy as a Loss of Prior Precision on Action Outcomes". bioRxiv. 2025-10-07T00:00:00.000Z. URL: https://doi.org/10.1101/2025.10.06.680685

[18] Sanchez-Aranda, L., de Geus, K., Solis-Urra, P., Sanchez-Martinez, J., Toval, A., Martin-Fuentes, I., Fernandez-Ortega, J., Alonso-Cuenca, R. M., Fernandez-Gamez, B., Olvera-Rojas, M., Coca-Pulido, A., Carlen, A., Moreno-Escobar, E., Garcia-Orta, R., Jann, K., Erickson, K., Esteban-Cornejo, I., Ortega, F. B.. "Cardiorespiratory fitness and cerebral blood flow in cognitively normal older adults and individuals with coronary artery disease: the AGUEDA and Heart-Brain projects". medRxiv. 2026-03-03T00:00:00.000Z. URL: https://doi.org/10.64898/2026.03.03.26347189

How this survey was made

This survey answers one question: “How do arterial spin labeling and other perfusion MRI methods relate regional cerebral blood flow changes to neuropsychiatric symptom trajectories in prodromal dementia cohorts?”

We keep our own collection of recent preprints and refresh it every day from bioRxiv, medRxiv, arXiv, PsyArXiv, SocArXiv and ChemRxiv. For each survey, we search that collection for recent work and keep the strongest matches: papers that use the same terms as the question, and papers on the same topic in different words.

The text was written from the titles and abstracts of those papers; the full texts were not read. Statements that draw on a particular paper are marked with a number in brackets. The reference list just above is the 18 papers whose work this text draws on, chosen from the 221 papers those searches returned; the rest of the shortlist is not printed. It is not a complete review of everything published on the question.

Treat this as a starting point and a reading list, not a peer-reviewed review. The searches behind this survey ran on October 22, 2026 and covered work from the previous two years.