Updated freezing of gait definition targets a long-standing measurement problem in Parkinson disease
Freezing of gait remains one of the most disabling motor complications of Parkinson disease and related disorders. It interrupts walking without warning, often during gait initiation, turning, or other high-demand situations, and it can leave patients vulnerable to falls, loss of confidence, and progressive dependence on caregivers. The central problem has never been the existence of the symptom itself. It has been the lack of a precise and widely accepted way to define it.
The new position paper from the International Consortium for Freezing of Gait attempts to close that gap by separating the phenomenon into two distinct definitions: a broad clinical definition for routine care and a more exact technical definition for video-based research scoring. In practical terms, that split recognizes what neurologists already face at the bedside. Patients need a clear description of what freezing is. Raters, meanwhile, need a stricter rule set for deciding exactly when an episode starts, ends, and qualifies for scoring.
Freezing is not a single uniform motor failure. During an episode, patients may stop entirely, take tiny ineffective steps, or show trembling of the legs while trying to move forward. Episodes are often triggered by cognitive load, sensory demands, emotional stress, or complex walking tasks such as turning and passing through a doorway. The heterogeneity matters because it makes the symptom difficult to capture with a single questionnaire item or a brief clinic walk.
A single sentence can hide a great deal of clinical complexity.
Why current assessments have struggled to produce consistent results
In everyday practice and research, patient-reported outcomes such as the new freezing of gait questionnaire and item 2.13 of the MDS-UPDRS are commonly used to identify freezing and estimate severity. Those tools are easy to administer, but they depend on recall, insight, and the patient’s willingness to describe a symptom that can carry embarrassment or shame. The paper warns that these measures can be biased by recollection errors and reduced disease insight, which weakens construct validity and creates measurement error. For small-to-moderate treatment effects, that limitation can be decisive.
Clinical observation offers another route, most often through item 3.11 of the MDS-UPDRS. That exam requires the patient to walk 10 meters, turn 180 degrees, and return. The protocol is useful for routine assessment, yet it may miss episodes that emerge only under stronger provoking conditions. Investigators have therefore built a range of gait tasks and measurement strategies, including 360-degree turning and doorway walking, to provoke freezing more reliably. The problem is that protocols have varied across studies, making direct comparison difficult.
This inconsistency has produced a measurement field with too many moving parts and too little agreement. Even when investigators rely on video recordings, the scoring rules have not been uniform. The most informative video outcome has been the percentage of task time spent freezing, or %TF, but the paper makes clear that the definition used to derive that number has remained uneven across studies. Inter-rater reliability for %TF has ranged from 0.50 to 0.99, a spread wide enough to affect both trial interpretation and confidence in the endpoint itself.
Video scoring and the gray zone between walking and freezing
The traditional definition of freezing, established at the 2010 workshop, described it as a “brief, episodic absence or marked reduction of forward progression of the feet despite the intention to walk.” The language is clinically familiar, but it leaves several unresolved questions. How brief is brief? What counts as markedly reduced progression? How should intent be judged from video alone? Where exactly does an episode begin, and where does it end?
Those ambiguities create a gray area. Two expert raters may look at the same sequence and disagree on whether a movement pause represents freezing, a gait hesitation, or a different abnormal gait pattern. The current workaround has been double independent scoring by two clinical raters, with a method for handling disagreements in borderline cases. The process improves rigor, but it is labor-intensive and costly. As a result, it is not used as often as it should be in clinical research.
The burden goes beyond workflow. When there is no dependable gold-standard reference, the validation of other freezing measures becomes less secure. That affects not only clinical scales and patient-reported tools but also automated detection algorithms designed to identify freezing objectively. If the reference standard is unstable, every downstream metric inherits some of that uncertainty.
A broader clinical definition and a stricter technical definition
The consortium’s main contribution is conceptual clarity. The clinical definition is intended to be concise and usable in patient care. The technical definition is designed for structured video annotation and research scoring. This distinction gives clinicians and researchers different tools for different tasks without forcing a single definition to do both jobs badly.
The working group also provided guidance for classifying people with Parkinson disease into subgroups with or without freezing of gait. That classification is more than a semantic exercise. Subgrouping affects study design, treatment comparison, and interpretation of therapeutic response. A trial enrolling a mixed population of freezing and non-freezing patients can dilute apparent treatment effects if the intervention works only in a specific subgroup.
The path toward the new definitions ran through the International Consortium for Freezing of Gait, formed after a 2022 review from a Scientific Panel of the International Parkinson and Movement Disorder Society identified the lack of standardization as a major obstacle to progress. The group met in Jerusalem in March 2023 and later in Atlanta in June 2025, with regular meetings in between, until consensus was reached. The paper also acknowledges the contribution of the late N. Giladi, whose work helped shape the field’s current framework.
Standardization is not a cosmetic change. It determines whether a treatment effect can be measured consistently across raters, sites, and studies.
What a harmonized definition could change in trials and daily care
A harmonized definition of freezing of gait should improve the reliability of %TF as an outcome measure and make comparisons across studies more meaningful. It may also strengthen the evaluation of clinical, patient-reported, and fully objective measures by giving them a more stable anchor. In the clinic, the updated clinical definition may help neurologists speak about freezing more consistently with patients and caregivers, especially when describing why a person freezes in some environments but not others.
The paper’s framework also supports the ongoing effort to validate a consensus-based freezing-provoking protocol, the Giladi Protocol, now being evaluated in 140 individuals with Parkinson disease in a large, diverse cohort. That work matters because a definition is only as useful as the test conditions used to elicit the phenomenon it describes. If the provoking task is inconsistent, even a well-written definition cannot fully solve the measurement problem.
The next challenge is implementation. Researchers still need to adopt the technical definition broadly, raters will need to apply it consistently, and trialists will need to decide whether the added precision offsets the practical burden of more detailed scoring. The paper offers a foundation, not a finished endpoint. Its real value will be measured in whether future studies can finally report freezing of gait with the kind of consistency the field has lacked for years.