What do body-first and brain-first Parkinson’s labels reveal?

Since the previous research note · Earlier expectation not tested

Track continuity

Since the previous research note: Why unexplained nerve damage is being studied for early Parkinson’s

Previous expectation: Cardiac nerve findings, sleep measurements and biological samples might help identify which people with unexplained peripheral nerve disease warrant closer neurological follow-up.

What this study adds: A separate Parkinson’s Progression Markers Initiative analysis links symptom-defined body-first and brain-first groups to subsequent clinical trajectories and adds imaging and genetic associations across prodromal and diagnosed participants. It does not test the earlier peripheral-neuropathy cohort or its proposed combination of measurements.

---

What changed in this study

A study published June 20, 2025, followed people with Parkinson’s disease and people showing specified risk features before a clinical diagnosis. It asked whether two proposed patterns of disease — body-first, beginning in nerves outside the brain, and brain-first, beginning in the brain — correspond to different clinical courses.

The answer was more informative than a simple division between two symptoms. Across clinical and prodromal groups, body-first participants generally had more motor and non-motor problems at the start and worsened more in several areas during follow-up. The study also connected the two groups with different spatial patterns in brain imaging and with different genetic associations.

Those findings make the labels potentially useful for organizing follow-up and research. They do not directly show where alpha-synuclein pathology first arose in any individual, or prove the physical route by which it spread.

Two labels built from symptoms

The researchers analyzed 1,120 clinically diagnosed participants and 910 prodromal participants from the Parkinson’s Progression Marker Initiative. The clinical group met standard diagnostic criteria and had evidence of dopamine-transporter deficits on imaging. The prodromal group had features associated with increased risk of later Parkinson’s, including reduced sense of smell, REM sleep behaviour disorder, or specified variants in Parkinson’s-associated genes.

The labels were assigned from baseline questionnaires rather than from a direct measurement of disease origin. Participants were classified as body-first when their scores for REM sleep behaviour symptoms, autonomic symptoms, or both were at least one standard deviation above the average of a healthy reference group. Autonomic symptoms are problems involving automatic body functions, such as the systems that regulate processes without conscious control. Participants whose scores for both measures remained within the reference range were classified as brain-first.

This operational definition matters. The labels were designed to reflect patterns that fit the proposed body-first and brain-first models, but they were proxies based on symptoms. They were not anatomical observations of the first site of alpha-synuclein aggregation.

The early difference widened during follow-up

At enrollment, body-first participants in both the diagnosed and prodromal groups had greater motor dysfunction, anxiety, and depression than their brain-first counterparts. A starting difference is not the same as a faster decline, so the researchers also modeled clinical trajectories over time and adjusted for factors including age, sex, education, and disease duration where appropriate.

Among people with diagnosed Parkinson’s, the body-first group had faster worsening in motor and non-motor measures. Axial symptoms and postural-gait problems progressed more quickly, as did treatment-related complications and depression. Global cognition and attention also declined more, and the body-first group had a higher risk of reaching the study’s threshold for Parkinson’s dementia, defined by a Montreal Cognitive Assessment score below 21.

The prodromal results showed a related pattern before a clinical diagnosis. Body-first participants developed more decline in motor function, including bradykinesia, rigidity, axial function, and postural-gait performance. Their attention, executive function, and memory also worsened more over time.

The prodromal body-first label was associated with a higher hazard of conversion to clinical Parkinson’s, with a hazard ratio of 2.02 and a 95% confidence interval from 1.15 to 3.55. The association remained after additional adjustments and after analyses intended to reduce the possibility that medication had hidden REM sleep symptoms. A hazard ratio compares the relative rate of an event between groups during the study period. It is not an individual’s absolute probability of diagnosis and does not say when diagnosis will occur.

A steadier classification may help with tracking

A clinical label is more useful for longitudinal work if it does not change simply because a symptom fluctuates or a person crosses an unstable threshold. In this dataset, body-first-to-brain-first changes were less frequent than the specified changes toward less severe categories in the cognition- and tremor-based classifications.

Using the entire baseline population as the denominator, the reported proportions shifting from body-first to brain-first were 8.6% at the three-year assessment and 6.7% at the five-year assessment. The corresponding proportions shifting from mild cognitive impairment to cognitively normal were 24.4% and 21.9%. The tremor-based comparison counted shifts from non-tremor-dominant to tremor-dominant categories.

This comparison used specified directions of category change; it was not an overall diagnostic-accuracy test. A lower rate of reclassification suggests that the body-like and brain-like symptom grouping may be steadier for observing progression, but it does not establish how well the labels would work as a population-wide screening system.

The researchers also used a small unsupervised deep-learning analysis to see whether clinical, imaging, and genetic measurements separated into similar patterns without imposing the body-first and brain-first definitions first. The resulting latent space showed regions that broadly aligned with the two phenotypes and with increasing clinical severity. Because it used a subset of the same research dataset, this result is an internal consistency check, not independent clinical validation.

Imaging showed different patterns across the brain

The imaging analyses gave the symptom labels a spatial counterpart, although each measure answered a different question.

One analysis examined functional connections involving the locus coeruleus, a brainstem region involved in sleep and autonomic functions. Body-first participants showed a pattern centered more on the caudal, or lower, part of the locus coeruleus. Brain-first participants showed a pattern more closely related to its rostral, or upper, part. Similar differences appeared in the prodromal groups. These findings are consistent with the proposed bottom-up and top-down models, but they remain associations with the clinical classification.

Dopamine-transporter imaging compared the more affected and less affected hemispheres. The more affected side had similar dopaminergic deficits in both phenotypes. The less affected side retained higher binding in brain-first participants, producing greater asymmetry between the two sides. The same kind of asymmetry appeared in a diffusion-based index of the glymphatic system, a brain fluid-clearance system assessed here through diffusion imaging. The less affected side showed higher values in brain-first cases, while the body-first group showed more impairment on that measure.

The asymmetry is a spatial pattern, not a photograph of disease traveling from one location to another. It also did not mean that every imaging measure tracked faster degeneration. In longitudinal dopamine imaging, the study found no significant difference between the phenotypes in the change of binding over time, and dopamine scans were not available for the prodromal participants. The glymphatic analysis also included a smaller subset than the main clinical comparison.

Genetic associations added another layer

The researchers examined 123 DNA variants previously associated with Parkinson’s disease to ask whether their distributions differed between the two phenotypes. After correcting for multiple comparisons, five variants showed phenotype-associated differences. Several were also linked with distinct outcomes among prodromal participants.

TRIM40 provides one example. Heterozygous TRIM40 alleles were associated with faster motor and cognitive decline in prodromal participants. IP6K2 provides a different example: people with two G alleles showed greater depression over the study period. These are separate kinds of association. A variant can be distributed differently between body-first and brain-first groups, and it can also be associated with a later outcome within a prodromal group. Neither observation by itself proves that the gene causes the phenotype or that changing its pathway would improve a patient’s course.

The study did not find a significant difference between the clinical phenotypes in carriers of GBA or LRRK2 mutations. The distribution of mutation carriers in the prodromal sample was also shaped by how that cohort was recruited, which included people selected for particular genetic or prodromal risk features. That design makes it difficult to treat the observed genetic pattern as a simple picture of the wider Parkinson’s population.

What these labels still cannot tell us

The study’s size and longitudinal design strengthen the comparison, and the findings remained in sensitivity analyses that considered comorbidities and other potential confounders. Even so, the classification depended on symptom questionnaires. Self-reported REM sleep symptoms can be affected by recall, awareness, or other sleep disorders. Autonomic symptom scores can vary with medication, coexisting conditions, and individual experience.

The researchers also lacked precise information about when sleep or autonomic symptoms began relative to motor symptoms. The authors warn that this could have placed some brain-first individuals with early REM sleep behaviour or autonomic symptoms in the body-first group. More objective sleep testing, autonomic assessments, cardiac imaging, and alpha-synuclein seeding assays could help characterize these groups more precisely, although no single test would by itself settle the anatomical origin of disease in an individual.

The genetic and imaging findings are similarly best read as converging group-level clues. The imaging mediation analysis linked glymphatic measures with part of the relationship between phenotype and motor impairment, but mediation is not proof of a causal pathway. The study also does not establish a calibrated prediction for an individual, independent usefulness in routine screening, or benefit from a treatment chosen according to one of these labels.

What clearer Parkinson’s trajectories could make possible.

For people at risk of Parkinson’s, a reproducible way to distinguish clinically meaningful trajectories could make follow-up more informative. For people already diagnosed, it could help researchers describe progression more precisely and match participants to trials testing treatments intended to slow disease. Those benefits would require the classification to hold up across different populations, with objective measures and prediction that work beyond this cohort.

A further possibility is treatment matched to the biology of a person’s disease, such as an intervention aimed at a mechanism more relevant to one progression pattern than another. That remains conditional on establishing actionable mechanisms and showing that targeting them changes outcomes. The immediate contribution of this study is narrower but important: it connects symptom-defined phenotypes with longitudinal clinical, imaging, and genetic differences that can now be tested more rigorously.

Key Sources

• Primary source: Clinical progression and genetic pathways in body-first and brain-first Parkinson's disease.
  DOI or Journal Page · Peer-Reviewed Paper · 2025-06-20