The research, led by Guillaume Carey of Maastricht University’s Mental Health and Neurosciences Research Institute together with colleagues in the Netherlands and France, drew on participants from the TRACK-PD cohort, a longitudinal study registered in the Dutch Trial Register. The team enrolled 105 non-demented people with Parkinson’s disease, of whom 31 experienced clinically significant anxiety and 74 did not, alongside 37 healthy controls. All participants underwent structural brain scanning on a 7 Tesla MRI scanner, a strength of the study because ultra-high-field imaging provides far greater anatomical detail than the 3 Tesla scanners used in most clinical settings.

The structures of interest were small but critically important: the amygdala, the hippocampus, the nucleus accumbens, the caudate nucleus, the putamen, and the thalamus. These nuclei sit at the heart of the brain’s circuitry for emotion, reward, and threat processing, and they are also heavily affected by the dopaminergic neurodegeneration that defines Parkinson’s disease. Because several of them are only a few cubic centimeters in size, reliably measuring their boundaries requires both high-resolution imaging and sophisticated computational segmentation, which is precisely where the new study pushed the field forward.

To extract quantitative measures from the scans, the researchers used FastSurfer, a deep-learning-based segmentation pipeline, to delineate each subcortical structure automatically. They then went beyond simple volume measurements by applying a shape analysis technique known as the Spherical Harmonics Point Distribution Model, or SPHARM-PDM. This method represents each three-dimensional structure as a smooth mathematical surface and identifies localized regions of inward or outward deformation. The approach matters because two structures can have identical overall volumes while differing dramatically in shape, and shape deformations can reveal where neurodegeneration or remodeling is concentrated within a nucleus.

The results were striking. Anxiety severity in the Parkinson’s patients was significantly associated with a bilateral reduction in caudate volume, meaning that the more anxious a patient was, the smaller the caudate nuclei appeared on both sides of the brain. The caudate, part of the striatum, is a major hub of the brain’s reward and motivation circuitry and is densely connected to prefrontal regions that regulate worry and fear. Its involvement fits with a growing understanding that anxiety in Parkinson’s is not simply a psychological reaction to a difficult diagnosis but may reflect degeneration in specific neural circuits.

Shape analysis added a second layer of evidence. Compared with non-anxious Parkinson’s patients, those with anxiety showed distinct shape deformations in the bilateral caudate, the bilateral thalamus, the left nucleus accumbens, and the left amygdala. When the anxious patients were compared with healthy controls, significant shape differences emerged in the left amygdala and left thalamus. The thalamus acts as a relay station funneling sensory and emotional information to the cortex, while the amygdala is the brain’s central threat-detection structure, and the nucleus accumbens anchors the reward system. Localized deformation in these nuclei suggests that anxiety in early Parkinson’s is accompanied by structural remodeling across a distributed emotional network rather than a single focal lesion.

One of the most important aspects of the study is its focus on early-stage disease. Most imaging research on psychiatric symptoms in Parkinson’s has examined patients in mid or late stages, when medication effects, motor complications, and widespread neurodegeneration make it difficult to isolate the biology of any single symptom. By restricting the analysis to non-demented patients in the early phase, the researchers could ask a cleaner question: does anxiety have its own structural signature before the disease has extensively reshaped the brain? The answer, based on these findings, appears to be yes.

The authors propose that these morphometric changes could serve as potential imaging markers for the occurrence and progression of anxiety in early Parkinson’s disease. Such markers would be valuable for several reasons. Anxiety in Parkinson’s is frequently under-recognized and under-treated, partly because patients and clinicians may attribute distress to the burden of the diagnosis itself. An objective neuroimaging correlate could help distinguish anxiety rooted in circuit degeneration from situational distress, guide screening for patients at highest risk, and provide a measurable outcome for clinical trials of new interventions targeting neuropsychiatric symptoms.

The study also carries caveats typical of cross-sectional designs. Because the data were captured at a single time point, the findings show association rather than causation; they cannot determine whether caudate atrophy precedes anxiety, follows it, or reflects a shared underlying degenerative process. The anxious group was also relatively small, with 31 patients, and longitudinal follow-up will be needed to establish whether the observed shape and volume changes predict the future course of anxiety symptoms. The authors and the TRACK-PD infrastructure are well positioned to address these questions as the cohort continues to be followed over time.

Nevertheless, the study represents a technical and conceptual step forward. It demonstrates that 7 Tesla MRI, combined with automated segmentation and harmonic shape modeling, can resolve subtle structural differences in deep brain nuclei that conventional imaging would likely miss. It reframes anxiety in Parkinson’s disease as a circuit-level phenomenon with detectable anatomical correlates, and it offers the research community a concrete set of candidate biomarkers centered on the striatum, thalamus, and amygdala. If future longitudinal work confirms these signatures, ultra-high-field morphometry could become a practical tool for identifying which patients with early Parkinson’s are most likely to develop anxiety, and for tracking whether new treatments protect both the motor and the emotional brain.

Subject of Research: Subcortical brain structure changes associated with anxiety in early Parkinson's disease detected by 7T MRI

Article Title: Alterations of subcortical structures are associated with anxiety in early Parkinson’s disease patients: a cross-sectional 7T MRI study

Article References: Carey, G., van Lier, M., Michielse, S., Wolters, A. F., Dujardin, K., Kuijf, M. L., & Leentjens, A. F. G. (2026). Alterations of subcortical structures are associated with anxiety in early Parkinson’s disease patients: a cross-sectional 7T MRI study. npj Parkinson's Disease. https://doi.org/10.1038/s41531-026-01532-3

Image Credits: AI Generated

DOI: 10.1038/s41531-026-01532-3

Keywords: Parkinson's disease, anxiety, 7T MRI, caudate nucleus, amygdala, thalamus, nucleus accumbens, subcortical structures, SPHARM-PDM, neuroimaging biomarkers, non-motor symptoms, TRACK-PD cohort