Parkinson’s disease is still diagnosed primarily at the bedside.

A clinician establishes the presence of parkinsonism, evaluates supportive features and red flags, considers alternative explanations, and may refine diagnostic confidence as the syndrome evolves.

By that stage, however, one or more disease-related biological processes may already have been active for years.

That has created one of the most important questions in modern neurology:

Can Parkinson’s disease be identified before its characteristic motor syndrome emerges?

The wording requires care.

A disease cannot be diagnosed before the relevant biological process exists.

The real objective is to identify Parkinson- or synucleinopathy-associated biology before conventional clinical diagnosis—during a phase in which molecular abnormalities, neurodegenerative changes, or prodromal manifestations may be present, but the defining motor syndrome has not yet emerged.

Several different concepts are therefore involved.

A person may have an increased genetic risk without having active disease.

Another may have prodromal manifestations but no demonstrable molecular biomarker.

A third may have detectable α-synuclein seeding activity while remaining free of a clinically defined neurodegenerative syndrome for an uncertain period.

And a person with isolated REM sleep behaviour disorder may eventually develop Parkinson’s disease, dementia with Lewy bodies, multiple system atrophy, or another clinically defined outcome within the synucleinopathy spectrum.

Risk, biological classification, prodromal disease, and clinical diagnosis overlap.

They are not interchangeable.

The biological idea began with the interval between the onset of disease-related biology and the appearance of clinically recognisable parkinsonism.

Under the Movement Disorder Society clinical criteria, parkinsonism is defined by bradykinesia together with rest tremor, rigidity, or both.

That establishes the core motor syndrome.

It does not establish when the relevant neurodegenerative biology began.

In many patients, the motor syndrome is unlikely to mark the beginning of that biology.

Clinical motor manifestations may be delayed by compensatory changes and functional reserve within the nigrostriatal system.

Striatal dopaminergic function can decline, neural networks can adapt, and pathological changes can accumulate before abnormalities become sufficient to produce clear clinical parkinsonism.

The relationship is not uniform.

Bradykinesia and rigidity correlate more closely with nigrostriatal dysfunction than tremor does, and not every patient develops tremor.

There is therefore no single biological threshold that applies identically to every patient or motor phenotype.

Researchers proposed that Parkinson-related disorders may pass through several broad stages.

First, molecular or cellular abnormalities may become detectable.

These could include seeding-competent α-synuclein, lysosomal impairment, mitochondrial dysfunction, altered proteostasis, inflammatory responses, or other processes that remain incompletely understood.

These abnormalities should not be assumed to occupy one fixed causal order.

Some may contribute to disease initiation.

Others may amplify progression.

Still others may be downstream consequences or compensatory responses.

Second, clinical features insufficient for a conventional diagnosis may emerge.

These can include polysomnography-confirmed isolated REM sleep behaviour disorder, hyposmia, constipation, autonomic dysfunction, subtle motor abnormalities, cognitive change, or mood symptoms.

Their predictive significance varies substantially.

Isolated REM sleep behaviour disorder is a strong marker of future synucleinopathy in specialist cohorts.

Constipation, depression, and nonspecific sleep disturbance are common in the general population and have much lower specificity.

Third, measurable biological or functional abnormalities may appear.

A seed-amplification assay may detect α-synuclein seeding activity.

Dopamine-transporter imaging may show reduced striatal binding consistent with presynaptic nigrostriatal dysfunction.

Digital assessments may detect subtle changes in movement.

These tests do not measure the same phenomenon.

A positive α-synuclein seed-amplification assay is evidence of a molecular seeding property.

Reduced DaT binding is evidence of dopaminergic-system dysfunction.

Neither directly measures the number of surviving neurons.

Finally, some individuals cross the clinical threshold for parkinsonism and may fulfil diagnostic criteria for Parkinson’s disease or another degenerative parkinsonian syndrome.

Longitudinal cohorts support the existence of a prolonged prodromal interval.

They do not establish one universal sequence of molecular, neuronal, and clinical events.

Different individuals may follow different trajectories.

Some may show α-synuclein seeding activity before measurable dopaminergic dysfunction.

Others, particularly within certain genetic forms of parkinsonism, may not fit neatly into an α-synuclein-first framework.

This biological heterogeneity has major implications for early diagnosis.

The therapeutic reasoning was nevertheless compelling.

If disease-related biology could be identified before extensive irreversible neuronal loss had occurred, a future treatment might be started during a more responsive biological window.

The aim would not merely be to assign a diagnosis earlier.

It would be to identify individuals at a stage when the disease trajectory might still be altered.

That possibility transformed early detection from a diagnostic ambition into a potential therapeutic strategy.

But earlier detection is valuable only if the information is sufficiently accurate and leads to an action that benefits the individual.

A marker may establish increased risk without proving active disease.

Another may establish α-synuclein seeding without predicting whether Parkinson’s disease or dementia with Lewy bodies will emerge.

A third may identify nigrostriatal dysfunction without revealing its cause.

Researchers therefore need markers capable of answering several distinct questions:

Is a disease-associated molecular abnormality present?

Is neuronal dysfunction present?

Is the process progressing?

Which clinical syndrome is likely to emerge?

How soon might phenoconversion occur?

And would intervention at this stage improve the outcome?

No single available marker answers all of them.

The first major evidence came from clinical observation.

Before Parkinson’s disease is diagnosed, many patients report symptoms that are not primarily motor.

Constipation, hyposmia, autonomic symptoms, mood changes, and sleep abnormalities may precede conventional motor diagnosis.

At a population level, these associations support the possibility of a prolonged prodromal phase.

They do not prove that any individual symptom was caused by Parkinson-related biology in a particular person.

Retrospective studies are also vulnerable to recall bias and to reinterpreting common past symptoms after a diagnosis has been made.

Prospective cohorts provided stronger evidence.

The most important clinical example is isolated REM sleep behaviour disorder.

During normal REM sleep, skeletal muscles are largely atonic.

In REM sleep behaviour disorder, recurrent dream-enactment behaviours arise during REM sleep and polysomnography demonstrates REM sleep without atonia.

Alternative explanations—such as medication effects, narcolepsy, seizures, or other sleep disorders—must be considered.

When the disorder occurs without an already defined neurological disease, it is termed isolated REM sleep behaviour disorder.

Longitudinal multicentre studies in specialist cohorts have shown that many individuals with isolated RBD later develop an overt neurodegenerative syndrome—most commonly Parkinson’s disease, dementia with Lewy bodies, or multiple system atrophy.

This makes isolated RBD one of the strongest recognised clinical markers of prodromal synucleinopathy.

It does not diagnose Parkinson’s disease specifically.

The eventual phenotype and timing remain uncertain for an individual.

Conversion estimates also depend on cohort composition and duration of follow-up.

Referral cohorts have historically included more men than women, and results may not generalise perfectly to community-detected disease or incidental REM sleep without atonia.

Hyposmia is common in Parkinson’s disease and can appear years before motor diagnosis.

But smell loss also occurs with ageing, sinonasal disease, viral illness, smoking, trauma, and several neurological disorders.

Its prevalence also differs among genetic subgroups.

Constipation may precede Parkinson’s disease but is highly prevalent in people who never develop a neurodegenerative disorder.

Orthostatic symptoms, urinary dysfunction, depression, anxiety, subtle cognitive changes, and minor motor abnormalities may contribute information, but none is sufficiently specific to diagnose prodromal Parkinson’s disease alone.

The difference between these markers is important.

Polysomnography-confirmed isolated RBD carries substantially greater predictive information than a nonspecific complaint of poor sleep.

A diagnosis of prodromal disease cannot therefore be built by simply counting symptoms.

The Movement Disorder Society research criteria for prodromal Parkinson’s disease formalised a probabilistic approach.

The criteria begin with an age-related prior probability.

Risk factors and prodromal markers are then incorporated using likelihood ratios within a naïve Bayesian model.

A person with isolated RBD, hyposmia, subtle motor signs, and abnormal dopaminergic imaging will therefore receive a much higher calculated probability than someone with constipation alone.

The framework was designed for research and for enrichment of observational or preventive-trial cohorts.

It does not establish an inevitable future diagnosis.

The label “probable prodromal Parkinson’s disease” corresponds to a calculated probability threshold, not pathological certainty or guaranteed phenoconversion.

The model also has limitations.

Its naïve Bayesian structure treats markers as though they were conditionally independent.

In reality, many prodromal features are biologically or statistically correlated.

Combining correlated markers may therefore produce probabilities that are not perfectly calibrated.

Performance also varies across cohorts according to age, recruitment method, available markers, and length of follow-up.

A second line of evidence came from dopaminergic imaging.

Dopamine-transporter SPECT measures radioligand binding to dopamine transporters within the striatum.

Reduced binding is consistent with presynaptic nigrostriatal dopaminergic dysfunction.

The measurement is influenced by terminal density, transporter regulation, medications, imaging acquisition, and analytical methods.

It should not be described as a direct count of dopaminergic neurons or terminals.

In clinical practice, DaT-SPECT can support the distinction between degenerative parkinsonism and selected non-degenerative mimics, such as essential tremor.

It cannot reliably distinguish Parkinson’s disease from multiple system atrophy, progressive supranuclear palsy, or other degenerative parkinsonian syndromes that also involve the nigrostriatal pathway.

In enriched research cohorts, reduced DaT binding may be detectable before overt parkinsonism and can increase the estimated risk of later phenoconversion.

It is generally considered evidence of established nigrostriatal dysfunction rather than the earliest molecular stage of disease.

It is neither sufficiently specific nor practical for general-population screening.

A normal study also does not exclude every case of very early clinically diagnosed Parkinson’s disease.

Genetics created another path toward presymptomatic identification, but the term “genetic Parkinson’s disease” conceals substantial biological diversity.

Pathogenic variants in SNCA, LRRK2, and VPS35 can cause autosomal-dominant parkinsonian disease, with penetrance and phenotype varying by variant.

Biallelic pathogenic variants in PRKN or PINK1 can cause recessive, usually early-onset parkinsonism.

A single heterozygous variant in a recessive gene does not automatically establish causal disease.

Some PRKN- or PINK1-associated disease may also lack typical Lewy pathology, illustrating that not all genetic parkinsonism belongs naturally within one α-synuclein-defined pathway.

Variants in GBA1 generally increase risk rather than make Parkinson’s disease inevitable.

Penetrance varies with the specific variant, age, ancestry, and additional genetic or environmental factors.

Genetic testing can therefore identify different states:

a causal genotype,

an incompletely penetrant susceptibility,

carrier status,

or increased population-level risk.

It does not necessarily demonstrate that active neurodegeneration has begun.

The most significant recent development has been the emergence of α-synuclein seed-amplification assays.

These assays do not simply measure total α-synuclein concentration.

They test whether a biological sample contains seeding-competent α-synuclein aggregates capable of promoting further aggregation under controlled laboratory conditions.

The process occurs in several steps.

A sample—most commonly cerebrospinal fluid—is added to a reaction containing recombinant α-synuclein substrate.

If seeding-competent aggregates are present, they promote conversion of the substrate into aggregated forms.

Cycles of fragmentation and elongation amplify the signal.

Fluorescent readouts indicate whether predefined evidence of seeding activity has been detected.

Assay platforms, reaction conditions, thresholds, and kinetic measurements are not identical across laboratories.

Large research cohorts, including the Parkinson’s Progression Markers Initiative, have shown high overall assay positivity among participants with clinically diagnosed Parkinson’s disease.

These studies have also detected α-synuclein seeding activity in some participants with isolated RBD, hyposmia, or genetic risk before conventional motor diagnosis.

This represented a major advance in detecting seeding-competent α-synuclein in living participants.

It did not create a universal Parkinson’s disease test.

PPMI is a deeply phenotyped research cohort rather than an unselected screening population.

Assay positivity varies among biological and genetic subgroups.

Some clinically diagnosed patients—particularly within selected LRRK2-associated or relatively preserved-olfaction groups—may have negative assays.

Conversely, a positive result identifies seeding activity associated with neuronal synucleinopathy.

It does not by itself establish whether Parkinson’s disease, dementia with Lewy bodies, or another phenotype will emerge.

It does not reliably predict when symptoms will begin.

Most existing assays are primarily interpreted qualitatively.

Whether kinetic measures can accurately predict phenotype or rate of progression remains under investigation.

Current evidence suggests that prognostic performance is less established than qualitative biological classification.

Assays also require greater harmonisation across platforms and laboratories before they can be interpreted as routine population-screening tests.

Phosphorylated α-synuclein has been detected in skin biopsies from patients with clinically established synucleinopathies.

Seeding assays and other α-synuclein measurements have also been investigated in skin and several other peripheral matrices.

These approaches are promising because they may eventually provide alternatives to cerebrospinal-fluid testing.

That remains conditional on adequate validation.

Sampling site, tissue handling, staining or amplification methodology, laboratory expertise, and interpretation can all affect performance.

Many studies have examined selected participants with established diagnoses.

Results from those cohorts should not be extrapolated directly to asymptomatic screening.

Wearable sensors, smartphones, speech analysis, typing patterns, gait monitoring, and sleep technologies can detect changes that may be subtle or intermittent during a clinic visit.

They can measure movement repeatedly and in real-world settings.

That makes them potentially valuable for monitoring progression or identifying enriched populations.

But an algorithmically detected movement abnormality is not equivalent to a diagnosis.

Digital measurements may be influenced by age, arthritis, fatigue, medication, comorbidity, device type, adherence, socioeconomic access, and differences between the populations used to train and apply an algorithm.

They require prospective validation against meaningful clinical and biological outcomes.

Taken together, these findings made earlier biological detection scientifically credible.

Longitudinal cohorts demonstrated an extended prodromal interval in selected groups.

Dopaminergic imaging identified preclinical nigrostriatal dysfunction in some individuals.

Genetics identified susceptible populations.

Seed-amplification assays detected seeding-competent α-synuclein before motor diagnosis in selected participants.

Digital technologies made subtle functional change measurable.

What the evidence did not establish was a single test capable of diagnosing future Parkinson’s disease accurately in every asymptomatic person.

The research has shown that selected biological and clinical signals can be detected before conventional Parkinson’s disease diagnosis in enriched populations.

It has not shown that future Parkinson’s disease can be diagnosed reliably in an unselected population before clinically defining manifestations appear.

That distinction is central.

The MDS prodromal criteria demonstrated that validated risk and prodromal markers can be combined into a formal probability estimate.

This gave researchers a consistent framework for cohort enrichment and longitudinal observation.

The criteria remain probabilistic by design.

A calculated probability is not equivalent to pathological confirmation, inevitable conversion, or a prediction of exactly when symptoms will appear.

External validation has been encouraging in some cohorts but variable in others.

Its performance depends on recruitment, marker availability, the correlation among markers, and the duration over which participants are observed.

Large multicentre studies of isolated RBD have provided some of the strongest evidence for a clinically recognisable prodromal synucleinopathy.

Over extended follow-up, many participants develop Parkinson’s disease, dementia with Lewy bodies, or multiple system atrophy.

Impaired olfaction, subtle motor change, cognitive abnormalities, autonomic dysfunction, and reduced DaT binding may further stratify risk.

However, phenoconversion is itself a clinical threshold.

It does not mark the biological onset of disease.

Not every participant converts during the observation period.

The eventual diagnosis varies, and precise individual timing remains difficult to predict.

The PPMI seed-amplification study provided strong evidence that seeding-competent α-synuclein can be detected in vivo in many clinically diagnosed and at-risk participants.

It also revealed biological heterogeneity within the clinical syndrome called Parkinson’s disease.

This supports—but does not by itself prove—the inference that clinically similar patients may have different underlying biological profiles.

A positive result supports the presence of α-synuclein seeding activity.

It does not by itself establish clinical Parkinson’s disease.

A negative result does not exclude every clinically diagnosed form, particularly in certain genetic or olfactory subgroups.

Longitudinal studies are now examining whether assay positivity or amplification kinetics predicts clinical progression.

Available evidence supports biological-classification value more strongly than individual prognostic accuracy.

It remains unclear whether the assay can reliably determine the future phenotype, time to conversion, or rate of decline in an at-risk person.

DaT imaging has shown that reduced striatal dopamine-transporter binding may precede overt parkinsonism in selected high-risk cohorts.

An abnormal scan can increase the estimated risk of subsequent phenoconversion.

But it is not disease-specific.

It does not identify α-synuclein pathology.

And because it generally indicates established nigrostriatal dysfunction, it may become abnormal after earlier molecular changes have already occurred.

Genetic studies have made it possible to identify some people with causal variants or elevated risk before motor manifestations develop.

These cohorts may eventually support trials designed to delay onset or prevent clinical progression.

But each genetic category requires separate interpretation.

SNCA multiplication, LRRK2-associated disease, biallelic PRKN disease, and GBA1-associated risk do not share the same inheritance, penetrance, pathology, clinical phenotype, or expected progression.

A genetic result therefore cannot be treated as a uniform diagnosis of presymptomatic Parkinson’s disease.

Cross-sectional skin-biopsy studies have reported encouraging detection of phosphorylated α-synuclein in patients with clinically established synucleinopathies.

These results support further development of peripheral tissue biomarkers.

They do not establish skin biopsy as a stand-alone test for asymptomatic people or for the general population.

Performance must be evaluated prospectively in earlier disease stages, unselected clinical populations, competing diagnoses, different laboratories, and diverse demographic groups.

Two influential research proposals attempted to organise this expanding biomarker evidence.

The neuronal α-synuclein disease integrated staging system, or NSD-ISS, proposes a biological definition and research staging framework based on α-synuclein seeding, neuronal-system dysfunction, and clinical-functional impairment.

It is a framework for neuronal α-synuclein disease.

It is not synonymous with every form of Parkinson’s disease or parkinsonism.

The SynNeurGe framework classifies participants across three dimensions:

synucleinopathy,

neurodegeneration,

and genetics.

The systems share the objective of moving research beyond symptom-defined categories, but they use different assumptions and organisational structures.

They should not be treated as interchangeable.

Both represent major conceptual developments.

Neither has been validated as a replacement for routine clinical diagnosis.

Important questions remain.

Should asymptomatic biomarker positivity be labelled “disease”?

How should clinically typical but SAA-negative Parkinson’s disease be classified?

Where do non-synuclein genetic forms of parkinsonism fit?

Can proposed stages reliably predict prognosis, clinical phenotype, or treatment response?

Will the systems generalise beyond highly characterised research cohorts?

These remain debated research frameworks rather than settled clinical standards.

The research supports several conclusions.

Selected clinical markers can identify groups at increased risk of future synucleinopathy.

Carefully validated multimodal combinations can improve risk stratification compared with isolated nonspecific symptoms.

α-Synuclein seed-amplification assays have substantially advanced molecular classification.

Nigrostriatal dysfunction can be demonstrated before overt parkinsonism in some high-risk individuals.

Genetic and biomarker-defined cohorts can support observational studies and future delay-of-onset trials.

But no existing approach can tell every asymptomatic person:

whether Parkinson’s disease will develop,

which syndrome will emerge,

when symptoms will begin,

how rapidly progression will occur,

or whether early treatment would improve the outcome.

That remains the gap between earlier biological detection and clinically useful diagnosis.

Every new biomarker teaches us something.

Few answer every question.

The future will belong not to the first positive test—

but to the first test that consistently improves patient care.

Parkinson’s research is increasingly incorporating biological classification alongside clinical diagnosis.

This shift has parallels with biomarker-based Alzheimer disease research, although the maturity, pathology, and clinical implications of the two fields are not equivalent.

The transition should therefore be approached carefully.

A clinical diagnosis identifies a syndrome that explains a patient’s current manifestations and guides management.

A biological classification identifies a molecular or physiological state associated with disease.

A risk estimate identifies increased probability without necessarily showing that active disease is present.

A prodromal classification indicates that features associated with future disease are already present, but conventional diagnostic criteria have not been met.

These categories overlap.

They should not be collapsed into one word.

Calling all of them “diagnosis” creates more certainty than the evidence supports.

Many possible prodromal features are common.

Constipation, depression, hyposmia, and sleep complaints occur frequently in people who will never develop Parkinson’s disease.

Screening an unselected population using nonspecific symptoms would therefore generate many false-positive or indeterminate classifications.

A future strategy may require sequential testing.

A broad and inexpensive assessment might identify an enriched population.

A more specific molecular or imaging test could then evaluate α-synuclein seeding or nigrostriatal dysfunction.

Longitudinal measurement might finally determine whether abnormalities are progressing.

This is a plausible model.

The optimal sequence, combination, thresholds, cost-effectiveness, and clinical consequences have not been established.

Not all Parkinson’s disease appears to follow the same route.

Some people may show detectable α-synuclein seeding activity before measurable dopaminergic dysfunction.

The precise ordering remains incompletely established.

Certain genetic subgroups may not conform to a synuclein-first framework.

Some individuals with α-synuclein-positive biology may develop a dementia-first, autonomic-predominant, or motor-predominant syndrome.

A useful biological classification may therefore need to describe both molecular pathology and the pattern of affected neural systems.

A biomarker result is more clinically useful when it provides information about timeframe.

A marker that hypothetically becomes positive decades before symptoms creates a different problem from one associated with conversion over the next few years.

For preventive trials, timing is critical.

Treat too late, and substantial irreversible neuronal injury may already have occurred.

Treat too early, and people may be exposed to years of cost, burden, and adverse effects without certainty that they would otherwise have become symptomatic.

At present, most available markers cannot provide sufficiently precise individual timing.

No treatment has been proved to prevent Parkinson’s disease in an asymptomatic biomarker-positive person.

This changes the ethical balance of testing.

Earlier knowledge may facilitate research participation, planning, general-health optimisation, and clinical observation.

Exercise and management of cardiovascular or general health remain reasonable recommendations, but they have not been proved to prevent phenoconversion in a biomarker-positive individual.

Testing may also generate anxiety, affect family relationships, alter a person’s perception of health, and carry employment or insurance implications depending on jurisdiction.

A biomarker result should therefore not be offered as though it were an ordinary laboratory value.

Pre-test counselling, informed consent, privacy, interpretation, and post-test support matter.

A clinically useful biomarker must do more than separate carefully selected patients from healthy controls.

It must demonstrate reproducibility across laboratories.

Pre-analytical handling must be standardised.

Assay thresholds must be calibrated.

False-positive and false-negative rates must be known in the population in which the test will actually be used.

Performance must be assessed across ancestry, sex, age, disease stage, genetic background, and competing diagnoses.

Most importantly, the result must add useful information beyond a careful clinical assessment.

If a biomarker is used as a surrogate endpoint in a prevention trial, change in that biomarker must reliably predict meaningful clinical benefit.

Most current Parkinson’s biomarkers have not yet met all these requirements.

So, can we diagnose Parkinson’s before it begins?

Not literally.

Nor can we diagnose future Parkinson’s disease with certainty in an asymptomatic person.

What we can increasingly do is detect selected clinical and biological signals before conventional motor diagnosis.

In enriched populations, isolated RBD can identify a high-risk prodromal synucleinopathy.

Seed-amplification assays can detect α-synuclein seeding activity.

Dopaminergic imaging can demonstrate nigrostriatal dysfunction.

Genetics can identify causal variants or elevated susceptibility in defined groups.

Each result answers a different question.

None provides the entire answer.

That is genuine progress.

It is not yet equivalent to a definitive early clinical diagnosis.

The most accurate conclusion is therefore measured.

Parkinson’s research is approaching an era in which biological detection may precede clinical diagnosis in selected individuals.

Whether those findings can predict an individual’s future accurately enough—and whether early intervention can improve that future sufficiently—to justify routine testing remains unknown.

The next advance may not be a single diagnostic test.

It may be a framework that combines clinical risk, molecular pathology, neuronal-system dysfunction, genetics, and longitudinal change.

The central question is no longer simply whether some Parkinson- or synucleinopathy-associated signals can be detected before motor diagnosis.

In selected populations, they can.

The harder question is whether we can interpret those signals accurately enough to help the person sitting in front of us.

—

Think Beyond the Headline.

Parkinson's disease gradually destroys the dopamine-producing neurons of the substantia nigra.

Modern therapies replace the missing dopamine—but not the neurons themselves.

Could regenerative medicine rebuild what neurodegeneration has taken away?

In the next issue of The Clinical Logic, we'll examine stem-cell therapies, neuronal transplantation, and whether replacing lost neurons could truly change the course of Parkinson's disease.