| Journal | Neuroscience |
| Study Type | Narrative Review |
| Population | Human participants and preclinical models |
Cognitive impairment is one of the most devastating and undertreated consequences of Parkinson’s disease, eventually affecting the majority of patients and ranking among the strongest predictors of caregiver burden and loss of independence. Conventional pharmacology offers limited traction. The endocannabinoid system is deeply embedded in the very neural circuits PD progressively dismantles, making it a scientifically credible and clinically urgent target for investigation. Understanding how cannabinoids interact with these circuits, and where the evidence currently stands, is essential context for any clinician navigating this conversation with patients.
Parkinson’s disease is far more than a movement disorder. While motor features dominate its clinical signature, the non-motor burden, including affective disturbance, sleep dysfunction, autonomic dysregulation, and above all cognitive impairment, accumulates relentlessly across the disease course and ultimately drives much of the disability patients experience. Cognitive decline in PD exists on a spectrum, ranging from subtle domain-specific deficits in attention, executive function, and visuospatial processing, to frank dementia, which eventually affects the majority of patients living with the disease for more than a decade. Despite this prevalence and severity, the therapeutic toolkit for PD-related cognitive impairment remains modest, creating genuine clinical pressure to understand whether emerging neuromodulatory systems might offer new leverage.
This narrative review from Kitchigina and Shubina, published online March 2026 in Neuroscience, synthesizes roughly three decades of accumulated evidence on how cannabinoids and the endocannabinoid system (ECS) interact with the neural substrates of cognition in PD. The ECS is a logical focal point for this inquiry. Endocannabinoid signaling is deeply embedded in the limbic circuits, corticostriatal networks, and hippocampal architecture that govern learning, memory consolidation, and executive function. CB1 receptors are expressed densely in the basal ganglia, prefrontal cortex, and hippocampus, precisely the regions most vulnerable to PD-associated neurodegeneration. CB2 receptors, expressed primarily on microglia and other immune-competent cells, modulate neuroinflammatory tone, which is now well established as a pathophysiological driver in PD, not merely an epiphenomenon. The ECS also participates in the regulation of synaptic plasticity through endocannabinoid-mediated retrograde signaling, influencing both long-term potentiation and long-term depression, the cellular mechanisms most closely associated with memory formation and cognitive flexibility.
The review spans preclinical rodent and primate models of PD, primarily those using neurotoxins such as 6-OHDA and MPTP, alongside the emerging but sparse clinical trial literature. In animal models, ECS dysregulation following dopaminergic denervation is well documented, and cannabinoid modulation has demonstrated neuroprotective, anti-inflammatory, and cognition-preserving effects across multiple experimental paradigms. THC, CBD, and synthetic cannabinoids have each shown capacity to influence memory, learning, and neuroinflammatory endpoints in these models, though the directionality and magnitude of effects vary considerably based on compound, dose, timing, and model specifics.
The clinical data tell a more complicated story. Trials that have included cognitive endpoints in PD populations using cannabinoid interventions are few, methodologically heterogeneous, and largely powered to assess motor or other non-motor outcomes. Cognition, when assessed at all, is typically a secondary measure, often using instruments that lack the sensitivity to detect domain-specific change or that conflate distinct cognitive processes under a single composite score. The variability in formulations studied, ranging from isolated CBD to full-spectrum products to synthetic CB1 agonists, further complicates any cross-trial synthesis. Dose ranges, treatment durations, and patient selection criteria have differed substantially, making it nearly impossible to draw unified conclusions about efficacy. The authors are appropriately candid about these limitations, calling for mechanism-informed trial designs with standardized, domain-specific cognitive endpoints as a prerequisite before any clinical recommendations can responsibly be made.
“What this review does well is map the conceptual architecture of why the ECS should matter in PD-related cognitive decline. The mechanistic rationale is genuinely compelling. We have a system with deep roots in the very circuits PD dismantles, a system capable of modulating neuroinflammation, synaptic plasticity, and neuronal excitability simultaneously. That is not a trivial convergence. In a disease where the pathological cascade is multifactorial and the cognitive damage is diffuse, a modulator that touches multiple nodes of that cascade has obvious theoretical appeal.
But the distance between a compelling mechanism and a defensible clinical recommendation is exactly where I find myself in practice, daily. The patients who ask me about cannabinoids for cognitive symptoms in PD are often thoughtful, well-informed, and genuinely motivated. They are not asking out of curiosity. They are asking because the conventional options have let them down, or they are watching a family member decline and looking for anything evidence-based they can offer. What I owe them is honest precision, not false hope and not reflexive dismissal.
The authors are right that the clinical database simply is not there yet. What has accumulated is a collection of trials that were not designed to answer the cognitive question, using tools that were not sensitive enough to detect cognitive change even if it existed. The heterogeneity is not just a methodological inconvenience; it reflects the fact that no one has yet run a trial that takes cannabinoid pharmacology in PD seriously as a primary research question with cognitive function as the primary endpoint. Until that happens, we are working with signal fragments.
What I watch for clinically in this population is whether cannabinoid use appears to be helping or harming sleep quality, anxiety, and pain, because these are the upstream contributors to cognitive performance that we can sometimes actually move. If a patient with PD is sleeping better, experiencing less resting anxiety, and managing pain with less reliance on opioids or sedating antihistamines, that secondary lift in daily cognitive clarity is real and worth tracking, even if it does not rise to the level of disease modification. That is not the same as treating the underlying neurodegeneration. But it is not nothing either.”
๐ง For clinicians managing PD, this review lands at an important inflection point. The endocannabinoid system is no longer a fringe consideration in neurodegeneration research. It is a recognized participant in the biology of PD pathology, and the field is now grappling with how to design trials rigorous enough to determine whether that biological participation translates into clinical opportunity.
In practice, several considerations are worth holding when a patient with PD raises cannabinoids in the context of cognitive concerns. First, current evidence does not support initiating cannabinoid treatment with the expectation of cognitive benefit. The preclinical data are genuinely promising but have not replicated in the clinic, and the absence of adequately powered, cognitively focused trials means that benefit remains unproven rather than disproven. These are not equivalent positions, and how a clinician frames this distinction matters considerably to an informed patient.
Second, cognitive function in PD is rarely impaired in isolation. Sleep disturbance, depression, anxiety, fatigue, and medication side effects all exert meaningful pressure on cognitive performance in this population. If cannabinoid use results in improvement in any of these comorbid dimensions, secondary cognitive effects may follow. This is a plausible mechanism for any subjective cognitive benefit patients report, and it should be part of the clinical conversation.
Third, the pharmacological complexity of cannabinoids in a PD context deserves careful individual assessment. THC, particularly at higher doses, carries well-documented risks for working memory impairment and psychomotor slowing, symptoms that overlap significantly with the cognitive profile PD patients are already navigating. CBD, while generally better tolerated cognitively, has its own pharmacokinetic considerations, including CYP450 interactions relevant to the medication burden many PD patients carry. Any discussion of cannabinoid use in this population needs to account for the specific compound, formulation, and dose rather than treating cannabis as a monolithic category.
Finally, if a PD patient chooses to pursue cannabinoid therapy, documenting baseline cognitive function with validated instruments and reassessing systematically over time is both clinically sound and contributes meaningfully to the real-world evidence base this field urgently needs.
๐ฌ Join the Conversation
Have a question about how this applies to your situation? Ask Dr. Caplan โ
Want to discuss this topic with other patients and caregivers? Join the forum discussion โ
Have thoughts on this? Share it: