The diagnosis of Parkinson's Disease (PD) is primarily based on the presence of cardinal motor features (Postuma et al., 2015), which develop due to degeneration of dopaminergic neurons in the substantia nigra. In addition, a variety of non-motors symptoms—caused by extranigral Lewy-body pathology and associated neuronal loss—also constitute an integral part of the disease, even in its earliest stages.

The prodromal phase of PD is mainly characterized by non-motor symptoms (NMS) such as hyposmia, constipation, REM sleep behavior disorder (RBD), and anxiety and/or depression (Mahlknecht et al., 2015). Except for RBD, these NMS are common in the general elderly population. Their association with incident PD has been extensively studied and is well established, leading to their inclusion in the Movement Disorder Society (MDS) research diagnostic criteria for prodromal PD (Berg et al., 2015). Among these NMS, isolated RBD has the highest specificity for future PD; in PD patients, the presence of RBD is associated with a more malignant phenotype, characterized by a higher burden of motor symptoms and NMS, including dysautonomia and cognitive impairment (De Pablo-Fernández et al., 2019).

Cognitive impairment is typically regarded as one of the principal NMS in the mid to late stages of PD and has been associated with reduced quality of life, loss of autonomy, and increased mortality (Foubert-Samier et al., 2020). Less is known about the types and incidence of cognitive abnormalities during the prodromal phase of PD, although cognitive changes have been reported to occur prior to diagnosis. Based on findings from three longitudinal studies (Darweesh et al., 2017a; Schrag et al., 2019; Weintraub et al., 2017), mild cognitive impairment (MCI) has been incorporated into the updated MDS criteria for prodromal PD (Heinzel et al., 2019). However, these studies are heterogeneous in many aspects, most notably in the assessment methods used and the baseline populations selected (enhanced risk vs. population-based).

In this study, we systematically reviewed prospective studies that report objective cognitive measures in individuals who were subsequently diagnosed with PD (i.e., individuals with presumed prodromal PD at baseline) to delineate the extent and range of cognitive changes occurring before the diagnosis of PD. We specifically focussed on studies conducted in population-based samples, as less is known in this setting, as opposed to studies in very high-risk cohorts or known prodromal populations, such as genetic mutation carriers or individuals with isolated RBD (Joza et al., 2023; Postuma et al., 2019). Furthermore, we conducted a meta-analysis to evaluate the association between lower baseline scores in a measure of global cognition and the incidence of PD.

The study was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Page et al., 2021) and was registered in PROSPERO (CRD42024547252).

This systematic literature review includes twelve prospective, population-based studies that examined objectively measured cognitive performance in individuals who were later diagnosed with PD—i.e. those presumed to be in the prodromal phase at baseline—compared to individuals who did not develop PD during follow-up (controls).

The findings are heterogeneous but collectively suggest that cognition may be mildly affected prior to diagnosis of PD. Of nine studies assessing global cognitive status, four found significant differences (Darweesh et al., 2017a,b; Pausch et al., 2016; Bock et al., 2023). This is supported by results of a meta-analysis showing that baseline MMSE scores were, on average, 0.3 points lower in future PD cases than in controls. Additionally, impairments appeared more frequently in tests of executive function and processing speed and attention, whereas deficits in visuospatial/visuoconstructive abilities and memory were less frequently observed. These domain-specific findings are substantiated by a recent comprehensive meta-analysis that also included cohort studies, such as those involving isolated RBD patients (Speelberg et al., 2022). This analysis found that poor executive function was associated with an increased risk of developing PD; in contrast, global cognitive impairment was not (Speelberg et al., 2022).

Several factors may explain the mixed results. First and most importantly, the included studies vary considerably in study populations and cognitive assessments. For example, in the Rotterdam study (Darweesh et al., 2017b), 7,386 participants were evaluated using nine tests covering global cognition, executive function, memory, processing speed, and attention. In contrast, the NEDICES study (Sánchez-Ferro et al., 2011) involved 115 participants assessed with the 37-item MMSE version. Second, cognitive changes may affect only a small proportion of individuals during the prodromal phase, similar to other prodromal NMS such as depression and anxiety (Jacob et al., 2010). Third, in affected individuals, cognitive changes may be subtle and therefore difficult to detect with the tests employed. Finally, different cognitive domains may be variably affected in individuals with incident PD, reflecting the heterogeneity observed in established PD.

The “dual syndrome” hypothesis postulates two broad profiles of cognitive impairment: (1) fronto-striatal dysfunction related to dopaminergic pathways degeneration, leading to impairments in executive function, working memory, and attention; and (2) posterior cortico-temporal dysfunction involving cholinergic pathways, resulting in visuospatial and recognition memory impairments (Brandão et al., 2020; Kehagia et al., 2013). According to this model, individuals with PD may develop or exhibit specific cognitive impairments depending on their patterns of pathology progression. In this literature review, executive function emerged as the most commonly impaired cognitive domain, consistent with experimental evidence suggesting earlier dopaminergic, rather than cholinergic, dysfunction in prodromal PD (Arnaldi et al., 2012; Hirano, 2021).

While most studies did not assess sex differences, one study focused on sex-specific cognitive trajectories before and after PD diagnosis (Bock et al., 2023). Women with incident PD showed no global cognitive decline, but showed a decline in executive function before diagnosis. Conversely, men with incident PD showed faster decline in global cognition before diagnosis and a prominent decline in executive function compared to women (Bock et al., 2023). This difference in global cognition may indicate some degree of sex-specific cognitive heterogeneity in the prodromal phase of PD. The executive function impairment observed only in men after diagnosis may be explained by the neuroprotective and neurotrophic effects of estrogen in women, which can also increase nigrostriatal dopaminergic activity (Bock et al., 2023; Bourque et al., 2019). It has been shown that women with higher lifetime exposure to estrogen have a lower risk of developing PD (Bourque et al., 2019), while those with lower lifetime estrogen exposure—such as earlier menarche (before age 12), later menarche (at age 14 or older), or a higher number of births—are associated with an increased PD risk.

Two of the twelve included studies also examined subjective cognitive complaints (Foubert-Samier et al., 2020; Darweesh et al., 2017a). Subjective cognitive complaints are known to carry a high risk of progression to neurodegenerative diseases, especially dementia. For instance, one longitudinal study assessed older participants with intact cognition or MCI at baseline, categorizing them based on the presence of self- or informant-reported complaints (Gifford et al., 2014). Among cognitively intact participants, those with any complaints had an increased risk of developing MCI or dementia. In MCI participants, only those with informant or combined (self and informant) complaints showed a higher risk of progression to dementia. Similarly, subjective motor complaints—such as stiffness or trembling—have been shown to increase the risk of future PD (De Lau et al., 2006; Schrag et al., 2015). Given the growing interest in cognitive impairment in PD, research has increasingly focussed on subjective cognitive complaints (Siciliano et al., 2024). Among the studies included in this review, two investigated this issue: the Rotterdam study (Darweesh et al., 2017a) found that subjective cognitive complaints appeared approximately 1.5 years before diagnosis, while the PAQUID study (Foubert-Samier et al., 2020) detected group differences between 12 and 4 years prior to diagnosis, but not after. The authors attributed these findings to depression, which can influence the individuals' perceptions about memory or cognition during PD development. Another cross-sectional study by Flores-Torres et al. (2021) reported more subjective cognitive complaints in individuals at high risk of developing PD (i.e. with one or more NMS) than in individuals without any prodromal PD features. A prospective six-years study by Schrag et al. (2023) found an association between subjective memory complaints (reported to general practitioners and recorded in a primary care database) and incident PD, although specific assessment tools or questions were not detailed. Taken together, these results highlight the potential of self-perceived cognitive difficulties as an early marker not only of AD, but of PD too. A recent systematic review and meta-analysis indicated a weak association between subjective cognitive complaints and objective cognitive performance in PD patients; however, patients who are cognitively healthy at baseline exhibit an increased risk of subsequent cognitive decline over an average follow-up of 3.16 years (Siciliano et al., 2024). In general, the predictive value of subjective cognitive complaints in neurodegenerative disorders remains under active investigation. While some studies suggest a link between early subjective complaints and future cognitive deterioration, others find no such association (Pitti et al., 2022; Reid and MacLullich, 2006), underscoring the need for further research. Notably, subjective cognitive complaints can also be assessed remotely, making them promising candidates for inclusion in cost-effective, internet-based screening batteries. Additionally, advances in device-based cognitive assessments, such as eye-tracking, may offer promising avenues for detecting changes related to neurodegenerative disorders while minimizing motor-related confounds (Bueno et al., 2019).

We should acknowledge limitations of the present literature review. First, there was considerable heterogeneity in the cognitive tests used (e.g., five different tests assessed language). Additionally, many studies relied on complex cognitive tests that simultaneously assess multiple domains, potentially obscuring domain-specific deficits. Second, the cognitive domains examined were unevenly represented, with most studies focussing on executive function, processing speed, attention, and memory, while fewer assessed language and visuospatial abilities. Third, some of the studies included had a substantial risk of bias. Fourth, there was a partial overlap between samples in two studies (Darweesh et al., 2017a,b). Due to the lack of precise information about this overlap, the final number of incident PD patients may not be completely accurate. Furthermore, the studies included did not specify whether participants had a genetic predisposition to develop PD or dementia. Fifth, most studies used the MMSE to assess global cognition, likely due to their population-based design and primary focus beyond neurodegenerative diseases. Despite the MMSE's broad screening nature, a statistically significant difference was observed in baseline MMSE scores between individuals with incident PD and controls. While a 0.3-point difference on the MMSE may not be clinically meaningful for single individuals, at a group level it provides valuable insights into early cognitive changes during the prodromal phase of PD. Specifically, these findings suggest that cognitive decline—detectable even through brief global assessments like the MMSE—may begin very subtly before more overt symptoms appear. In this context, more sensitive screening tools, such as the Montreal Cognitive Assessment (MoCA) or the Parkinson's Disease Cognitive Rating Scale (PD-CRS), may offer greater efficiency in detecting subtle cognitive changes in prodromal PD (Brandão et al., 2023; Yang et al., 2023). According to a recent review (Bezdicek et al., 2025), a comprehensive neuropsychological assessment—including tests of memory, visuospatial functions, and language—is recommended for the early identification of cognitive alterations in association with PD. Finally, our review primarily focused on objective cognitive measures. A section dedicated to subjective cognitive complaints was added retrospectively after noticing that two of the studies included also addressed this aspect. It is important to note that “subjective cognitive complaints” is just one of several terms used to describe this issue. Future research should consider the variability of terminology to facilitate a more comprehensive understanding of the significance of subjective cognitive changes in the prodromal phase of PD.

In conclusion, cognitive impairment may be present during the prodromal phase of PD. A comprehensive neuropsychological evaluation targeting the most affected cognitive domains, along with subjective cognitive complaints, could enhance early detection. Further prospective studies are needed to delineate the trajectory of cognitive changes in prodromal PD and determine their significance for risk screening and the subsequent clinical course of individual patients.