Parkinson’s disease (PD) is characterized by distinct motor dysfunctions; bradykinesia, tremors, and rigidity, that are associated with degeneration of dopaminergic neurons in the substantia nigra leading to a loss of dopamine signaling (Kordower et al., 2013; Kouli et al., 2018; Wirdefeldt et al., 2011). Importantly, non-motor symptoms occur in roughly 90% of PD patients (McDowell and Chesselet, 2012) with a prodromal period including depression, anxiety, and cognitive difficulties preceding motor symptoms (Durcan et al., 2019). Notably, up to 50% of PD patients experience a form of mild cognitive impairment characterized by subtle cognitive changes not always apparent on standardized tests (Goldman and Litvan, 2011). These changes may directly affect quality of life including through therapeutic non-compliance (Straka et al., 2019; Daley et al., 2012). Some have proposed that cognitive deficits can be present up to 10 years before PD diagnosis (Fengler et al., 2017). Because the subtleties in cognitive dysfunction in PD are under-detected and under-treated in clinical practice, it is highly important for skilled clinicians to recognize changes in cognitive domains that are risk factors for early decline. To aid in such recognition, a need exists for techniques combining decision-making behavior with psychophysiological correlates, such as eye tracking, which may enhance the characterization and early diagnosis of PD (Antoniades and Spering, 2014).

One cognitive domain particularly vulnerable to early PD-related decline is executive functioning. Executive functioning is essential for goal-directed behavior and adjustments to novel situations (Nejtek et al., 2021; Friedman and Miyake, 2017). The prefrontal cortex (PFC) is critically involved in executive functioning, such as through selection and inhibition (Aron et al., 2004; Dulas and Duarte, 2014; Dulas and Duarte, 2016), conflict monitoring (Botvinick et al., 2004) and risk and reward processing (Bechara et al., 1994; Bechara and Damasio, 2005). In line with this, individuals with PD often exhibit impairments in evaluating reward-based outcomes (Perugini et al., 2018), they display a tendency toward risk taking, and they struggle to learn from negative feedback (Colautti et al., 2021).

The Iowa Gambling Task (IGT) is one of the most widely used assessments for understanding decision making abilities under conditions of reward and uncertainty in individuals with PD (Colautti et al., 2021; Salvatore et al., 2021). Salvatore et al. (2021) propose that the IGT may be sensitive to detecting prodromal changes due to its close relationship with several key regions involved in the neurological dysfunction of PD patients (e.g., frontal cortex and locus coeruleus). Many studies have reported significantly lower IGT performance in PD patients compared to controls (e.g., Salvatore et al., 2021; Castrioto et al., 2015; Xi et al., 2015), perhaps reflecting an inability to balance reward and risk (Kobayakawa et al., 2010; Kobayakawa et al., 2017). Neurobiologically, these deficits may stem from a substantial loss of dopaminergic function in the midbrain (e.g., substantia nigra and ventral tegmental area) and noradrenergic dysfunction in the locus coeruleus, both of which have extensive projections to the prefrontal cortex (Salvatore et al., 2021; Alberico et al., 2015; Mather and Harley, 2016). Accordingly, some studies have estimated a 40–77% loss of dopamine-based innervation in the ventral tegmentum and a 63% cell loss in norepinephrine-based cell loss in locus coeruleus in patients with PD (Alberico et al., 2015; Mather and Harley, 2016).

Given that eye movements are closely linked to both dopamine and norepinephrine signaling status, eye tracking offers a promising and noninvasive window into the neural mechanisms underlying cognitive decline in PD. Fixations, saccades, and blinks are robust reflections of both perceptual and cognitive functions (Spering, 2022), and multiple mid-brain and frontal connections are involved in eye movements (McGinty et al., 2016; Ding et al., 2009; Hui and Beier, 2022). For instance, dopaminergic activity involved in learning, memory, and goal-oriented behavior can be indirectly measured through spontaneous eye blink rate during visual exploration (Van Slooten et al., 2019). Specifically, dopamine is thought to modulate the frequency of these spontaneous eye blinks involved in reward-driven behavior and cognitive flexibility (Jongkees and Colzato, 2016). Eye blinks are also related to dopaminergic activity in basal ganglia by contributing to both motor and cognitive functioning (Willett et al., 2023; Redgrave et al., 2010; Eckstein et al., 2017; Zhou et al., 2021). Understanding how cognitive deficits manifest through eye movement abnormalities may provide valuable early indicators of neurodegeneration (Wong et al., 2018). Unfortunately, research linking decision-making deficits with eye movement abnormalities in early PD remains limited (Wong et al., 2018; Archibald et al., 2013; Wang et al., 2016; Tsitsi et al., 2021; Ranchet et al., 2017). To our knowledge, no study has integrated the IGT with eye tracking to examine decision-making in individuals with early-stage PD. This novel approach may provide deeper insight into the neurocognitive mechanisms underlying PD-related cognitive decline and could offer a valuable tool for early disease detection.

The first goal of the present study was to test motor functioning, verbal fluency, impulsivity, and cognitive flexibility in a sample of early-stage PD patients evaluated during their “ON” state compared to age-matched and education-matched healthy controls. Our second goal was to record eye movements during completion during decision-making using the IGT. We first hypothesized that participants with PD would perform worse on motor tasks and measures of attention and cognitive flexibility compared to healthy controls. Second, we hypothesized that the number of fixations, fixation durations, and number of eye blinks, would significantly differ between individuals with PD and healthy controls during the IGT. Finally, we hypothesized that overall behavioral performance on the IGT would be worse for our PD group.

We secured Institutional Review Board approval through the University of North Texas on February 26th, 2024, prior to beginning this study (Reference Number IRB-23-425). Participants were recruited through community-based and non-profit organizations, healthcare providers, local businesses, universities, and word of mouth. Prospective participants were given a brief overview of the motor tasks and a general overview of the cognitive assessments used. Examinations occurred in the Neurocognitive Laboratory at the University of North Texas and were directly observed by the primary investigator with the length of each assessment ranging from 45 min to 150 min depending on clinical status of the participant.

We compared early-stage to mid-stage individuals with PD to age-matched healthy controls using a novel combination of the IGT coupled with eye tracking and additional cognitive and motor tests. As expected, PD participants experienced higher adverse impacts on daily living and more motor complications than matched healthy controls. Individuals with PD also exhibited significantly longer completion times on both Trail Making Tests, significantly worse decision-making over time in the IGT, and a significant reduction in blink rate despite similar eye fixation and saccade rates.

In accordance with our first prediction, participants with PD scored higher on each of the three UPDRS subscales than controls, and completion times on our measure of attention and cognitive flexibility were significantly longer in the PD group compared to controls. Additionally, participants in the PD group exhibited a higher number of errors with increased task demands compared to their control counterparts. This pattern is consistent with established norms (Tombaugh, 2004) and supports previous findings of impaired executive function in PD (Aron et al., 2004; Dulas and Duarte, 2014; Dulas and Duarte, 2016; Kudlicka et al., 2011). Our measures of impulsivity and verbal fluency revealed similar scores between groups, therefore it is possible that our PD sample had not yet reached diagnostic thresholds for impulse control disorders (Giovannelli et al., 2023) or impoverished oral semantic fluency (Hedman et al., 2022).

With respect to our prediction about IGT performance, we did observe an overall difference in net scores between groups that approached significance. This is in accordance with pattern in the previous literature of diminished decision-making in PD patients compared to healthy controls (Castrioto et al., 2015; Xi et al., 2015; Kobayakawa et al., 2010; Kobayakawa et al., 2017). Upon closer examination of decision-making across task blocks (Kobayakawa et al., 2017), we discovered a unique pattern. By the fifth and final block of trials, individuals with PD had exhibited a net negative score (i.e., losses), whereas the healthy controls had accumulated a positive score (winnings), and this difference was highly significant. The current literature on IGT in PD presents mixed findings regarding behavioral performance (Salvatore et al., 2021). Some studies report impaired performance in PD compared to healthy controls, whereas others do not (e.g., Poletti et al., 2012). Our findings underscore the importance of evaluating decision-making changes over time within a task, rather than relying solely on aggregate performance. The fact that PD participants maintained a disadvantageous deck selection strategy by the task’s end, potentially reflects insensitivity to aversive outcomes in reward-based learning (Salvatore et al., 2021; Kobayakawa et al., 2010; Kobayakawa et al., 2017). This aligns with the hypothesis that dopaminergic and noradrenergic changes in PD, particularly involving the substantia nigra, basal ganglia, and prefrontal cortex functionally manifest as an insensitivity to negative feedback, thus impairing error correction.

With respect to our eye movement predictions, we observed a highly significant 70% reduction in eye blinks during decision-making in PD participants. This is in accordance with prior research documenting reduced task-independent spontaneous blink rates in mid-stage PD patients (Iwaki et al., 2019; Kimura et al., 2017). A substantial reduction in blink rate may reflect disrupted dopaminergic and norepinephrinergic function, which plays a role in both motor control and cognitive processing (Van Slooten et al., 2019; Jongkees and Colzato, 2016; Willett et al., 2023; Redgrave et al., 2010). Additionally, decreased dopaminergic activity could contribute to impaired motor execution in regions innervated by the corticobulbar tract. This pathway involves regulating movements of the face, head, and neck by influencing the cranial nerves, including those that control the extraocular muscles (cranial nerves III, IV, and VI). While this pathway does not directly regulate eye blinks, its interaction with brainstem regions could indirectly impact voluntary eye movements. Notably, a lower blink rate may contribute to the characteristic hypomimia, or “facial masking,” commonly observed in PD, which has psychosocial implications for both patients and caregivers (Bianchini et al., 2024). Importantly, these findings support the utility of blink rate as a diagnostic marker, particularly during cognitive tasks such as the IGT in the early stages of PD (Van Slooten et al., 2019; Tao et al., 2020). Researchers should reconsider the exclusion of eye blinks as artifacts during eye-tracking analyses given their strong association with dopaminergic function. Contrary to our expectations, we did not find significant group differences in the average number of saccades, fixations, fixation duration, or pupil dilation. However, exploratory analyses revealed significant correlations between fixation duration and decision-making performance in the middle and final blocks, suggesting a potential cognitive interaction that warrants further investigation.

This study is the first, to our knowledge, to integrate eye tracking with the IGT in individuals with PD. Our findings reveal nuanced cognitive and physiological distinctions that may serve as early indicators of disease progression. Subtle strategy shifts in decision-making over time, coupled with distinct eye movement patterns, suggest that a combined approach may enhance diagnostic sensitivity in early PD. Future research should prioritize multimodal assessments that integrate cognitive testing (including decision-making tasks as well as working memory and executive functioning tasks) with non-invasive physiological measures like eye tracking. Such an approach could provide a more comprehensive characterization of cognitive dysfunction in PD and facilitate earlier detection of subtle impairments. This early detection may have valuable practical implications. For instance, a recent review by Salvatore et al. (2021) highlights that the neurobiological changes in dopaminergic and noradrenergic innervation at the earliest stages of PD lend themselves for interrogation by the IGT in detecting early PD pathology and predicting treatment adherence. Notably, non-strategic decision-making in PD may correlate with non-compliance or self-discontinuation of therapeutic regimens. Given that up to 70% of PD patients opt to discontinue treatment despite life-threatening risks (Straka et al., 2019; Daley et al., 2012), these findings highlight a critical deficit in evaluating long-term consequences in decision-making. Our results also suggest that reduced eye blinks may represent a prodromal symptom that can be objectively measured before the onset of profound motor symptoms. Furthermore, it is possible that a reduction in eye blinks, facial masking, and apparent apathy are interconnected. Understanding the relationship between these symptoms is crucial for understanding the progression of PD.

Our study is not without limitations. Participants were a well-educated, high-functioning, and racially/ethnically homogeneous subset of the population with access to medical care. Future research should aim for greater demographic diversity in sample recruitment. Additionally, our relatively small sample size may have limited statistical power to detect group differences in saccadic behavior and fixation patterns. Our brief measure of impulsivity may also have lacked sensitivity; therefore, future studies should incorporate PD-specific assessments of impulse control disorders (Giovannelli et al., 2023). Future work should strive to balance sex of patient and control samples. Previous literature suggests that healthy men tend to perform better than healthy women on the task (Zanini et al., 2024), however we are not aware of any normative sex data on IGT performance of individuals with PD.

Finally, we suggest that creating a standardized version of the IGT including a pre-stimulus baseline adapted for eye tracking would enhance the precision of oculomotor analyses. The current study was limited by the short duration of each trial and the variability of self-paced responses, which restricted fine-grained analyses of discrete eye movement phases (e.g., initial presentation, deck selection, feedback).

In conclusion, the IGT holds promise for assessing cognitive changes in Parkinson’s disease, particularly for evaluating decision-making abilities. As the disease progresses, subtle changes in cognition, such as executive function, decision-making, and risk assessment, may decline. This has key implications for daily life and treatment decisions. Considering the currently projected doubling of individuals who will develop the disease in their lifetime (Obeso et al., 2017), there is valuable utility in combining cognitive assessments with physiological markers to gain a more comprehensive understanding. Integrating sensitive measures such as eye tracking with a range of cognitive tests, including decision-making, working memory, and other executive function tasks, alongside other disease markers may provide a simple yet powerful method for detecting dysfunction early in disease progression. This multimodal approach could improve characterization of disease progression and enable earlier intervention before significant motor symptoms emerge.