◉ Colloquy — research, out loud
medRxiv · 2026 · doi:10.64898/2026.08.24.26361220
Background and Objectives: To test whether the positive correlation between rest tremor (RT) ipsilateral striatal DAT binding reflects a tremor-specific mechanism or simply globally better-preserved dopamine terminals in RT patients. Methods: We compared contralateral correlations binding, tremor, bradykinesia rigidity two cross-sectional Parkinson disease cohorts from Parkinson's Progression Markers Initiative (baseline, N=1055; follow-up, median 2.2 years, N=652). tested survived permutation testing that shuffled severity scores isolating severity-dependent effects group-level used out-of-sample prediction to compare how well predicted symptom presence versus severity. Results: Bradykinesia showed large correlations, distinguishable null every comparison, both Rest tremor's was not most comparisons. Striatal ipsi contralateral, its (AUC 0.55-0.61) but (R2<0.002), cohorts. Conclusions: Our findings argue against direct pathophysiologic link amplitude dopaminergic function. More parsimoniously, binding-RT likely distinct degeneration pattern patients with rather than graded, dose-dependent circuit mechanism.
Researcher A Rest tremor in Parkinson disease has a weird relationship with dopamine. The more dopamine loss you see on one side of the brain, the more tremor shows up on that same side — which sounds like it should make sense, but it probably doesn't. That's what this paper challenges.
Researcher B Wait, that sounds backwards. Parkinson tremor is supposed to come from the opposite side of the brain, right? The left hemisphere controls the right side of the body.
Researcher A Exactly. That's the catch. The authors are saying this ipsilateral correlation — same-side binding and tremor — might not reflect how tremor actually works. It might just be a statistical mirage.
Researcher B So why are we confused about this in the first place?
Researcher A Because the ipsilateral tremor-dopamine correlation keeps showing up. Multiple groups have found it, with different imaging tracers, in different cohorts. It's reproducible. But reproducible doesn't mean you understand what it means.
Researcher B What are the two competing ideas?
Researcher A One is that there's a real, direct circuit: less dopamine on the left side directly causes more tremor on the left side — a dose-dependent relationship. The other is simpler: patients who get tremor just happen to have a different overall pattern of dopamine loss across the whole striatum, and that difference alone creates the correlation without any tremor-specific mechanism.
Researcher B So it's not about the tremor circuit at all — it's just a marker that tremor-prone patients have different dopamine loss overall?
Researcher A That's the hypothesis they're testing. And the key insight is: these two ideas make different predictions about severity.
Researcher B How do you tell them apart?
Researcher A If dopamine loss directly causes tremor amplitude, then more dopamine loss should predict more tremor. If it's just a group-level difference, dopamine loss should predict whether someone has tremor at all, but not how bad it is.
Researcher B Clever. So what data did they use?
Researcher A Two cohorts from the Parkinson's Progression Markers Initiative, or PPMI. The baseline cohort had 1,055 patients, average age 63 and a half years. The follow-up cohort had 652 patients, measured about 2.2 years after enrollment. Each patient had a dopamine transporter PET scan — that's DAT-SPECT — and a motor exam called the MDS-UPDRS-III done within 91 days.
Researcher B And they measured rest tremor, bradykinesia, and rigidity separately?
Researcher A Right. They pulled out left and right severity scores for each symptom from the exam. Then they did something clever: a permutation test. They held patients without tremor fixed, then randomly shuffled the severity scores among tremor-positive patients 10,000 times, and asked whether the original correlation was any different from that shuffled null.
Researcher B That's a clever way to isolate severity-dependent effects from just group differences.
Researcher A Exactly. They also did out-of-sample prediction — leave-one-out cross-validation — to see whether dopamine binding predicted tremor presence versus severity. They looked at both the caudate and putamen, both ipsilateral and contralateral.
Researcher B Okay, so what happened?
Researcher A Bradykinesia and rigidity behaved as expected. They showed large contralateral correlations with striatal dopamine binding — in the baseline cohort, correlation coefficients between minus 0.20 and minus 0.38, all highly significant. And those correlations survived the permutation test, meaning they're real severity-dependent effects.
Researcher B The opposite side of the brain, as it should be.
Researcher A Right. Rest tremor was different. It showed modest positive ipsilateral correlations — plus 0.11 to plus 0.22 — but those correlations were not distinguishable from the permutation null. In other words, when you shuffle severity around, the correlation doesn't go away.
Researcher B So the correlation was there, but it wasn't a real severity effect.
Researcher A Exactly. Then they tested prediction. Rest tremor presence was predicted from dopamine binding on either side — ipsilateral or contralateral — with an AUC of 0.55 to 0.61. But severity among tremor-positive patients? R-squared of 0.002 or less. Essentially zero.
Researcher B Whereas bradykinesia did the opposite?
Researcher A Yes. Contralateral putamen binding predicted both presence and severity of bradykinesia — AUC 0.65 to 0.74 for presence, R-squared 0.015 to 0.056 for severity. That's the pattern you'd expect from a direct, dose-dependent mechanism. The same pattern held in the follow-up cohort.
Researcher B So the methods work — they can detect real dose-dependent effects — but tremor doesn't show them.
Researcher A Correct. The quieter finding is that tremor presence was predicted equally well from ipsilateral and contralateral binding, which is weird if the mechanism is ipsilateral.
Researcher B What are the limitations?
Researcher A The authors flag a few. Symptom severity came from ordinal MDS-UPDRS-III items, which coarsen the data. If there's a real but very small dose-response relationship, ordinal scores might miss it. But the same coarse measure detected the expected relationship for bradykinesia and rigidity, so that's not a complete escape.
Researcher B What else?
Researcher A Bradykinesia is measured from five items with a wider dynamic range than tremor, which is measured from fewer items. That asymmetry could matter. The authors themselves say the presence analysis for bradykinesia should be interpreted cautiously — the severity analysis is more informative.
Researcher B And beyond what they flag?
Researcher A The sample is from a single multi-site cohort — PPMI. Replication in independent cohorts would strengthen the findings. Also, this is a preprint, so it hasn't been peer-reviewed yet. And the permutation test, while clever, is more statistically demanding than conventional testing, so it might be conservative.
Researcher B Fair. But the fact that their methods caught the expected effect for bradykinesia and rigidity suggests they're not just too conservative.
Researcher A Right. That's a good internal control.
Researcher A Three groups should care. First, movement disorder neurologists and Parkinson researchers. This changes how to interpret the ipsilateral tremor-dopamine correlation they've been seeing. It's not a circuit mechanism, so don't build models around ipsilateral dopaminergic control of tremor.
Researcher B What's the practical implication?
Researcher A If you're developing therapies targeting dopamine for tremor, this suggests you should think about what makes tremor-prone patients different globally, not about local ipsilateral circuits.
Researcher A Second, people working on Parkinson subtypes. Tremor-dominant PD follows a distinct clinical course from akinetic-rigid PD. This paper suggests there's a real, distinct dopaminergic phenotype in tremor-dominant disease — not just a difference in symptom presentation, but a difference in the underlying pattern of dopamine loss.
Researcher B That matters for prognosis and stratification.
Researcher A Exactly. Third, computational modelers and biostatisticians. This is a nice example of how a robust, reproducible correlation can be misinterpreted without the right statistical test. The permutation approach and the distinction between predicting presence versus severity are both generalizable ideas.
Researcher A The full citation: Mendonça, Marcelo D, and Joaquim Alves da Silva. Ipsilateral rest tremor-dopamine transporter correlation reflects a broader dopaminergic difference, not tremor-specific pathophysiology. medRxiv, 2026. DOI: 10.64898, slash, 2026.08.24.26361220.
Researcher B The thread is open on Colloquy.