Increasing Contrast Improves Object Perception in Parkinson's Disease with Visual Hallucinations

ABSTRACT

Background

Deficits in basic vision are associated with visual hallucinations in Parkinson's disease. Of particular interest is contrast sensitivity loss in this disorder and its effect on object identification.

Objectives

Evaluate whether increased contrast improves object perception in persons with Parkinson's disease and visual hallucinations, without dementia.

Methods

We assessed 26 individuals with mild to moderate idiopathic Parkinson's disease, half of whom reported one or more episodes of hallucinations/unusual perceptual experiences in the past month, with a letter‐identification task that determined the contrast level required to achieve 80% accuracy. Contrast sensitivity was further assessed with a chart that presented stimuli at multiple spatial frequencies. The groups were closely matched for demographic and clinical characteristics except for experience of hallucinations.

Results

Relative to participants without visual hallucinations, those with hallucinations had poorer spatial frequency contrast sensitivity and required significantly greater contrast to correctly identify the letters on the identification task. Specifically, participants with hallucinations required a mean contrast of 52.8%, whereas participants without hallucinations required 35.0%. When given sufficient contrast, the groups with and without hallucinations were equally accurate in letter identification.

Conclusions

Compared to those without hallucinations, individuals with Parkinson's disease and hallucinations without dementia showed poorer contrast sensitivity. Once contrast was individually enhanced, the groups were equally accurate at object identification. These findings suggest the potential of visual perception tests to predict, and perception‐based interventions to reduce, hallucinations in Parkinson's disease.

Among the most distressing of the non‐motor symptoms of Parkinson's disease (PD) are visual hallucinations (VH; reviewed in ¹ , ² ). Reported prevalence of VH ranges from 16% to 75%. ³ , ⁴ , ⁵ , ⁶ Efforts to understand this visuoperceptual phenomenon include the characterization of minor hallucinations (eg passage hallucinations, presence hallucinations, and illusions) and their onset before PD dementia. ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ Researchers have identified numerous risk factors for VH in PD without dementia including exposure to dopaminergic medications, older age, disease severity and duration, comorbid depression, and altered visual processing. ⁴ , ⁵ , ¹⁵ , ¹⁷ , ¹⁸ Altered visual processing is likely to be an important contributor (reviewed in ¹⁹ , ²⁰ , ²¹ ). Those with PD who report VH typically show reduced visual acuity, contrast sensitivity, and color discrimination, as well as compromised visuospatial and visual object perception. ²² , ²³ , ²⁴ , ²⁵ , ²⁶

Structural imaging studies have found gray matter loss in multiple cortical and subcortical brain areas associated with visual processing in PD with VH ( ²⁷ , ²⁸ , ²⁹ , ³⁰ ; reviewed in ³¹ , ³² , ³³ ). Functional imaging studies of PD‐VH without dementia have also provided evidence supporting the bottom‐up model of visuoperceptive processing, including reduction of activation, perfusion, or metabolism in the visual association cortices during rest or visual stimulation tasks. ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ Others have found less activation of the lateral occipital‐temporal cortex in PD with VH while the participants viewed pictures of objects gradually appearing out of random noise. ³⁹ , ⁴⁰

Studies using novel visual behavioral paradigms have shown impaired ambiguous object perception in PD‐VH ⁴¹ that could be normalized by exogenous cues. ⁴² These studies did not systematically examine the role of basic vision in object identification in PD‐VH, however, or the interaction between lower‐level visual processes and higher‐order processing of objects. There are numerous reports of deficient contrast sensitivity in PD (reviewed by ²⁷ , ³³ ), as well as of the ability of individuals with PD without dementia to benefit from increased contrast as they engage in object recognition tasks such as letter identification ⁴³ and visual search of numbers. ⁴⁴ The question is raised as to whether increasing contrast between the background and targeted object ameliorates object recognition deficits among individuals with PD without dementia who experience hallucinations. We hypothesized that persons with PD and VH would have poorer contrast sensitivity than those without VH and therefore would require more contrast to identify target objects; but when sufficient contrast was provided, the groups would demonstrate comparable accuracy.

Method

The Boston University Institutional Review Board approved the study. Participants provided written informed consent according to the Declaration of Helsinki.

Participants

Participants included individuals with idiopathic PD (U.K. Parkinson's Disease Society Brain Bank diagnostic criteria;⁴⁵ ) who were recruited from the Boston University Parkinson's Disease Clinic and Movement Disorders Center at the Boston Medical Center and through local support groups. They underwent screening by phone or, if in the lab for a concurrent study, in person. Screening consisted of asking them the first five questions from the Boston University Hallucinations and Unusual Perceptual Experiences questionnaire (BU‐HUPE, described below. See Appendix S1 for the questionnaire). Those who endorsed any item (occurrence in the past month) were included in the visual hallucinations group (VH) and were then administered the rest of the questionnaire. Those who endorsed no items were included in the non‐ visual hallucinations group (NVH) and did not proceed with the questionnaire. Participants were interviewed about their medical history to rule out confounding diagnoses (eg stroke, head injury, serious medical illness, ocular disorders) and brain surgeries that would have influenced performance on the measures of interest. All underwent detailed neuro‐ophthalmological examination at the New England Eye Institute to exclude potential contributory comorbidities such as cataract, glaucoma, and macular degeneration.

Participant characteristics are provided in Table 1. Our sample size of 13/group reflected the difficulty of recruiting individuals with mild–moderate PD who would endorse concurrent VH and also met all inclusion and exclusion criteria (ie no dementia diagnosis). The NVH group was drawn from a larger pool to closely match the VH group in regard to demographics and clinical characteristics. The VH and NVH groups were matched for age and education, and number of women ⁶ and men. ⁷ All were cognitively healthy as indexed by scores on the modified Mini‐Mental State Examination (mMMSE;⁴⁶ with a cut‐off score of 27 upon conversion to standard MMSE) and by general neuropsychological assessment that was conducted for parallel studies, including tests of memory, executive function, language, and visuospatial function. No significant differences were found between the groups in MMSE scores (t [24] = 0.78, P = 0.44), nor in depressive symptoms as measured by the Beck Depression Inventory‐II (t [24] = 0.74, P = 0.47). ⁴⁷

TABLE 1.

Participant characteristics

	PD‐VH (n = 13)	PD‐NVH (n = 13)	P‐value
Age	65.5 (10.1)	67.5 (5.5)	0.54
Education, years	17.3 (1.7)	17.7 (2.3)	0.63
MMSE	28.3 (1.3)	28.6 (0.9)	0.44
BDI‐II	7.6 (6.7)	6.0 (4.2)	0.47
Disease Duration, Years	5.3 (2.5)	5.9 (3.9)	0.63
UPDRS Total	32.1 (9.3)	27.3 (12.9)	0.29
LED (mg/day)	387.3 (224.5)	296.3 (220.3)	0.34

All values are reported as means (standard deviations). PD‐VH = Parkinson's disease with visual hallucinations; PD‐NVH = Parkinson's disease without visual hallucinations; MMSE = Mini‐Mental State Examination; BDI‐II = Beck Depression Inventory‐II; LED = levodopa equivalent dose. PD‐VH and PD‐NVH did not significantly differ on any demographic or PD clinical characteristic assessed.

Disease severity was determined with the Unified Parkinson's Disease Rating Scale (UPDRS, Sections 0‐3 ⁴⁸ ). The VH and NVH groups did not significantly differ on the UPDRS Total Score (t [24] = 1.08, P = 0.29; Table 1). All were in Hoehn & Yahr ⁴⁹ stages 1 to 3 indicating mild to moderate disease. All were taking medication for their parkinsonian symptoms and the groups did not differ in the mean levodopa equivalent dosage (t [21] = 0.97, P = 0.34).

Visual hallucinations were assessed with the 32‐item self‐report BU‐HUPE questionnaire, which probed for experiences during the past month including simple hallucinations (eg light flashes, sparks, stars, or geometric shapes), illusions (eg misperceptions), vivid sense of presence, sense of movement in periphery (passage), and complex hallucinations (eg people, animals, objects), as well as temporal and descriptive characteristics. BU‐HUPE was developed in our lab after careful consideration of the questionnaires and structured interviews available in the literature. The questionnaire included items from the University of Miami Parkinson's Disease Hallucinations Questionnaire, ⁵⁰ the North‐East Visual Hallucinations Interview, ⁵¹ the Queen Square Visual Hallucination Inventory, ⁵² and the Vision Questionnaire designed for PD. ⁵³ , ⁵⁴ Participants were informed that the types of visual experiences in PD differ significantly from those experienced with psychiatric disorders (eg schizophrenia, psychosis), in order to reduce the possibility that they would hide such symptoms if they were in fact experiencing them.

Materials and Procedures

Tests were given binocularly. The examiner was not blind to the participant's status regarding hallucinations. The contrast sensitivity tests are standard in the field and unlikely to be subject to examiner bias, as they require only that the forced‐choice response be recorded on a test notepad (spatial frequency contrast sensitivity; three choices) or keyboard (contrast sensitivity letter identification; four choices). For the latter test, the luminance levels were set by the computer algorithm (staircase procedure), not by the examiner.

Near Acuity

A Snellen eye chart (Lighthouse, Long Island City, New York) was given at 16 inches to the corner of the eye, the same distance as for the contrast sensitivity chart described next, under the same lighting conditions. Participants read a series of progressively smaller letters starting at the top row and moved down until they could no longer identify one half of the letters on that row. Standard Snellen scores (minimal angle of resolution) were log‐transformed to perform group comparisons.

Spatial Frequency Contrast Sensitivity

The Functional Acuity Contrast Test (FACT) was used to assess static, near spatial frequency contrast sensitivity. ⁵⁴ The FACT chart was viewed in a chin rest from a distance of 16 inches. The lighting for the chart was within the recommended luminance of 68–240 cd/m². The chart displayed a 9 by 5 array of gratings, with each grating's diameter subtending 1.7° of visual angle. Contrast decreased monotonically in nine steps from left to right with a range of 0.602 to 2.255 (0.59%–25% Michelson contrast), and a log step increment range of 0.109 to 0.176 (SD = 0.014). There were five spatial frequencies, including 1.5, 3, 6, 12, and 18 cycles per degree (cpd), moving down the columns. The gratings were oriented either vertically, tilted 15° to the left, or 15° to the right. The task was to indicate verbally or by hand posture the direction in which the lines were oriented. A contrast level was determined for each spatial frequency by finding the minimal perceptible contrast level needed to correctly identify the orientation of the grating for a given row.

Contrast Sensitivity, Letter Identification

We used a design similar to that described in Amick et al. ( ⁴³ ; Fig. 1). Participants were briefly presented with four letters (H, O, T, and X) on a monitor screen of a Mac G3 computer, viewed from a distance of 16 inches. Each of the four letters was 0.475 inches in height and subtended 1.7° of visual angle. These letters were displayed within a box measuring 256 × 256 pixels, which functioned as a background and was held at a constant gray level. Stimuli were presented one at a time on the screen for 12 ms followed by an interstimulus interval of 59 ms at the constant gray level, and then followed by a visual mask for 506 ms. The visual mask consisted of the overlapping letters H, O, T, and X that covered the entire 256 × 256 pixels box. The luminance of the target letters varied using a staircase procedure that established the luminance level required to obtain 80% target identification accuracy (ie contrast threshold). A 2.2 gamma function was used to relate gray level to display luminance (minimum luminance = 19.2 cd/m²; maximum = 82.5 cd/m²). Contrast levels were calculated using the Michelson contrast formula, (max Lum ‐ min Lum)/(max Lum + min Lum), where max Lum of the target stimulus was the threshold obtained from the third subtest and min Lum was the luminance of the contrast background, which remained constant throughout the subtests.

FIG 1.

Letter identification task. The letter identification task used an interleaving staircase procedure that identified the target contrast threshold that the participant needed to correctly identify the letter at an 80% accuracy rate. Each of the four letters H, O, T, and X was displayed within a box measuring 256 × 256 pixels, which functioned as a background and was held at a constant gray level. Stimuli were presented one at a time on the screen for 12 ms followed by a constant interstimulus interval of 59 ms, and followed by a visual mask for 506 ms. The visual mask consisted of the multiple overlapping letters H, O, T, and X positioned randomly in the display field to create an unstructured pattern (more unstructured than represented in this schematic for clarity). The participants' task was to orally report the perceived letter. See text for details.

Participants were dark adapted for 5 minutes and the task was performed in a darkened room. Once adapted, the participant was to name out loud the letter flashed on the screen, while the examiner recorded the verbal responses by keyboard presses corresponding to the four letters (H, O, T, and X).

The letter identification task was divided into four subtests, and the only parameter that changed was the luminance level of the target letter. In the practice subtest, participants received 20 trials and the target letter was presented at the maximum contrast of 91%. This subtest ensured that the participant understood and could perform the task reliably. The second subtest served as a second practice to orient the participant to the process of threshold measurement. This task used an interleaving staircase procedure that identified the required target contrast threshold the participant needed to correctly identify the letter at an 80% accuracy rate. The stopping criterion in the threshold estimation was a standard error of 20%. The actual threshold estimate, and the one used for group difference analyses, was determined on a third subtest that used a stricter stopping criterion of a 15% standard error to once again determine the target contrast level required for participants to achieve an error rate of 20%. The fourth subtest consisted of 20 trials presented at the participant's final threshold level to ensure that the threshold estimate was reliable. This subtest provided the numbers of errors participants made at their contrast threshold.

Data Analysis

For all vision tests, the groups were compared using independent‐samples t‐tests. For exploratory analyses with smaller VH subgroups, Mann Whitney U tests were used. On the FACT chart, one NVH participant could not read the chart at 12 or 18 cpd, and one VH participant could not read it at 18 cpd. These are the highest spatial frequencies, at which sensitivity decreases most at older ages. On acuity, one NVH participant had missing data. On letter identification, the main test of interest, the hypothesis was that relative to the NVH group, the VH group would show poorer contrast sensitivity, translating into the need for significantly higher contrast in order to attain 80% accuracy on the letter identification task. For this latter task, contrast threshold at the third subtest of object identification was the dependent variable. Outliers were identified in each group. Contrast thresholds above or below two standard deviations from the group mean were eliminated from the analysis, resulting in one NVH participant being eliminated. Two VH participants had missing data for number of errors in the letter identification task; their threshold data were included but not error data.

Results

Visual Hallucinations: Characteristics

Types of hallucinations. The 13 participants with PD who reported VH during the past month endorsed a mix of types of hallucinations/unusual perceptual experiences including complex, simple, and minor (passage, presence, illusion). Six reported complex hallucinations (one with solely complex hallucinations; two with additional simple hallucinations, perception of movement, and vivid sensation of presence; two with vivid sensation, sensation of movement, and illusion, and one with brief sensation of movement). Five reported simple hallucinations (two with additional perception of movement, two with perception of movement and vivid sensation of presence, one with illusions). Two participants reported only brief sensations of movement (passage). Eleven of the 13 reported that their visual hallucinations occurred during the afternoon, evening, or night, whereas only one reported them in the morning while waking up; one chose not to answer this question. Three participants noted the visual experiences in dim lighting conditions. For all participants, hallucinatory images did not make any noise, did not move, were of normal size and solid, and tended to appear suddenly. Seven participants reported that they realized that the percepts were not real, which denotes preserved insight; five chose not to answer this question. Four participants reported recent changes in their medication treatment, but only two related their medication change to changes in the characteristics of the hallucinations.

Frequency and duration. One participant experienced VH daily, three weekly, four every 2 weeks, and four once, with one person not answering the frequency question. Six reported their hallucinations lasting up to one second; two a few seconds; four from 15 to 60 seconds, and one a few minutes.

Vision Tests

The groups did not differ on corrected near acuity. Mean logMar (and standard deviation) for VH was 0.16 (.11) (20/29 Snellen) and for NVH was 0.23 (0.19) (20/34 Snellen) (t [23] = 1.15, P = 0.26).

Spatial Frequency Contrast Sensitivity

A mixed design ANOVA with five levels of spatial frequency (1.5, 3, 6, 12, and 18 cpd) and two levels of group (VH, NVH) was performed to examine differences in contrast sensitivity. Results revealed a significant main effect of spatial frequency (F [4, 88] = 118.6, P < 0.001, partial η² = 0.84), a significant main effect of group (F [1, 22] = 12.51, P < 0.002, partial η² = 0.36), and a significant interaction between spatial frequency and group (F [4, 88] = 3.0, P < 0.05, partial η² = 0.12). Planned t‐tests were performed to examine between‐group differences. Sample size was 13/group except NVH n = 12 at 12 and 18 cpd, VH n = 12 at 18 cpd. The VH group showed significantly poorer contrast sensitivity than the NVH group at all spatial frequencies except at 1.5 cpd, for which there was still a trend in this direction (1.5 cpd: t [19] = 1.8, P < 0.09; 3 cpd: t [24] = 2.1, P < 0.04; 6 cpd: t [24] = 3.0, P < 0.007; 12 cpd: t [15.21] = 4.0, P < 0.001; 18 cpd: t [22] = 2.3, p < 0.03).

Contrast Sensitivity, Letter Identification

The entire PD sample required a mean contrast of 44.3% (SD = 19.3) to perform the letter identification task at 80% accuracy, with a mean total error score of 5.4 (SD = 3.6). The VH and NVH groups significantly differed in the amount of contrast required to perform the task. Sample size was 13/group except NVH n = 12 for thresholds and errors, VH n = 11 for errors. The VH group required a mean contrast of 52.8% (SD = 21.0%; range 33.7%–90.8%) and the NVH group 35.0% (SD = 12.2%; range 14.5%–69.0%) (t [23] = 2.6, P < 0.02, partial η² = 0.22) (Fig. 2). At the adjusted contrast levels, the two groups did not significantly differ in the number of errors made at the criterion level of accuracy (t [21] = 0.08, P = 0.94). The VH group made on average 5.5 (SD = 3.7) errors, and the NVH group made 5.3 (SD = 3.7).

FIG 2.

Mean contrast thresholds on the letter detection task at the criterion accuracy of 80%. Error bars represent standard error of the mean. The entire group (left column) required an overall mean contrast of 44.3% (SD = 19.3). Relative to participants with PD without visual hallucinations (PD‐NVH, n = 12), those with hallucinations (PD‐VH, n = 13) required significantly more contrast to correctly identify the target letters, *P < 0.05. At the adjusted contrast levels, the VH group performed at the same level of accuracy (i.e., comparable number of errors) as the NVH group. See text for details.

Within the VH group, those with complex VH (n = 6) required 61.4% contrast (SD = 18.2%; range 42.8%–90.8%), whereas those with simple hallucinations (n = 5) and those reporting only sensations of passage (n = 2) required 45.3% (SD = 25.1%; range 22.3%–78.6%) and 46.0% contrast (SD = 17.3%; range 33.2%–58.2%), respectively. Those with complex hallucinations made on average 6.2 errors (SD = 4.5), whereas those with simple or passage hallucinations made 5.3 (SD = 3.1) and 3.5 errors (SD = 2.1), respectively. Because of the small numbers and resultant reduction of power, statistical analysis of the VH subgroups must be considered exploratory. The subgroups (complex, n = 6; simple/passage, combined because of their similar means, n = 7) did not differ in contrast threshold (U = 30.0, SE = 7.0, P = 0.23) or errors (U = 17.5, SE = 5.3, p = 0.66). Those with simple/passage hallucinations did not differ from the NVH group in contrast threshold (U = 52.5, SE = 11.8, p = 0.38) or errors (U = 27.5, SE = 9.4, p = 0.80), whereas those with complex hallucinations required significantly more contrast than the NVH group (U = 65.0, SE = 10.7, p = 0.005); number of errors did not differ (U = 39.5, SE = 10.6, p = 0.75). The finding that there were no group or subgroup differences on number of errors indicates that once the contrast level was adjusted, all (sub)groups were able to perform object identification at the same level.

Discussion

We examined whether enhanced contrast aided the recognition of visual stimuli among individuals with PD who experienced visual hallucinations (VH) in the absence of dementia. Our hypothesis was that relative to persons with PD without VH, those with VH would show poorer contrast sensitivity and would require more contrast to correctly identify target objects (letters). Consistent with our hypothesis, the group with VH showed poorer contrast sensitivity across multiple spatial frequencies, and on the letter identification task required a mean contrast of 52.8% to reach criterion, whereas the NVH group required only 35.0% contrast. At the adjusted contrast levels, the VH group performed at the same level of accuracy (ie comparable number of errors) as the NVH group. The results indicated that persons with PD with VH had deficient contrast sensitivity, but once the target stimulus was adjusted to their individual contrast threshold level, they were able to accurately identify the object. Our results also showed that it was the participants with complex VH who drove the significant VH‐NVH group difference in contrast thresholds.

Brain imaging studies have shown structural and functional changes in regions typically associated with visual processing among those with PD and hallucinations (reviewed in ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² , ³⁵ , ³⁶ , ³⁸ ). These studies, coupled with behavioral designs corroborating impaired object recognition in this population, ³⁹ , ⁴⁰ , ⁴¹ have served as the basis for the inclusion of impaired visual processing (both low‐level and high‐order visual regions) in theoretical models of visual hallucinations (reviewed in ¹⁹ , ³³ ). Our findings further suggest that the enhancement of the salience of visual characteristics in the ambiguous world of hallucinating PD individuals may improve cognitive performance in this population.

Our study provides empirical evidence that object perception can be enhanced by the manipulation of low‐level basic visual processes in PD with VH. The results are consistent with our research showing improvement in object identification with contrast enhancement in individuals with PD unselected for hallucinations status compared to healthy matched adults. ⁴³ , ⁴⁴ Amick and colleagues ⁴³ used an interleaving staircase procedure similar to that used in the current study and found that the PD group performed normally on the object identification task when the contrast of the target stimulus was enhanced to compensate for the low‐level vision deficit of the participant. The PD group required a mean contrast of 42.9% whereas the normal control group required a mean contrast of 25.5% to perform the task at the criterion error rate. Toner and colleagues ⁴⁴ used a similar interleaving staircase procedure in order to examine the role of enhanced contrast on a visual search task in young adults, healthy older adults, and persons with PD and found no difference in their ability to search and detect the targets once the contrast level was adjusted to their individual contrast threshold. The PD group required a mean contrast of 47.9% and the healthy age‐matched sample a mean contrast of 31.7%. In our present sample, the mean contrast combined across VH and NVH, 44.3%, accords with the thresholds reported in the studies by Amick et al. ⁴³ and Toner et al. ⁴⁴ Our lab has also evaluated the effect of low‐level visual cues on the resolution of perceptual ambiguity and found that individuals with PD without dementia and healthy older adults benefited from the low‐level visual cues. ⁵⁵ The results of the current study raise the possibility that visually‐targeted interventions may ameliorate the perceptual abnormalities experienced by those with PD and VH.

We found that there was greater variability in the amount of contrast required for accurate letter identification in the VH group (SD of 21.0%) than the NVH group (SD of 12.2%). Individuals who reported complex hallucinations, as opposed to solely simple or minor hallucinations, required higher contrast. Minor hallucinations have received increasing attention as potential early clinical markers of more complex non‐motor symptoms. ⁹ , ¹¹ A recent review by Lenka and colleagues ¹¹ suggested that minor hallucinations (passage, presence, and illusions) could be early markers of larger‐scale neurocognitive network dysfunction, cognitive impairment, well‐defined visual hallucinations, depression, and REM sleep behavior disorder. In our sample, two PD individuals reported brief sensations of movement (passage), six reported a combination of simple and minor hallucinations, and five of the six PD individuals with complex VH also reported simple or minor hallucinations including passage or presence hallucinations and illusions. If low‐level visual processing deficits represent an early marker of risk of developing complex VH in the future, it will be important to specifically categorize visual phenomena, including hallucination type, early in the course of PD. ⁷ , ¹¹ , ¹² , ⁵⁶

The present study was limited by small sample size. In order to examine the ability of cognitive performance (letter identification task) to benefit from sensory enhancement, we focused on individuals without a dementia diagnosis, although it is known that VH often co‐occurs with dementia in PD. ⁴ , ⁵⁷ Our sample size reflected the difficulty of recruiting individuals with mild–moderate PD who would endorse concurrent VH in the absence of dementia. Some potential participants who did experience VH may have chosen not to endorse them, as these symptoms are thought by the general populace to be associated with psychosis and schizophrenia. Moreover, some participants showed reluctance to endorse certain items on the hallucinations questionnaire. For example, five of the 13 did not answer the question about whether they realized that the percepts were not real. It is noteworthy that despite the small samples, the VH and NVH groups still significantly differed in contrast sensitivity on both measures.

Because we limited our sample to persons with PD without cognitive impairment, we are restricted in our ability to generalize our results to those with dementia. There are many studies of PD manifesting with both hallucinations and dementia. It was our intention in this study to focus on PD without dementia in order to determine whether contrast sensitivity loss occurred in this cohort, which would suggest that perceptual impairment may arise early and may in fact underlie the development of hallucinations. The absence of dementia eliminates this factor as a potential explanation of performance on perceptual and cognitive tasks in our and others' studies.

Strengths of the study include the focus on the effect of contrast enhancement on object identification in addition to documenting contrast sensitivity loss in PD; careful matching of the VH and NVH groups on demographic and clinical characteristics known to be associated with hallucination risk (age, PD severity and duration, depression, dopaminergic medications); and the use of established psychophysical methods not only to measure contrast sensitivity but also to enhance it in the effort to improve performance. A further contribution is our inclusion of a detailed VH questionnaire categorizing subtypes (passage, presence, illusions, simple and complex hallucinations), with subtypes being relevant to the study results.

In conclusion, persons with PD who are cognitively healthy, without dementia, may nevertheless have contrast sensitivity loss that is associated with the experience of visual hallucinations and with difficulties in object identification. Prospective, longitudinal studies are needed to examine whether contrast sensitivity loss or other perceptual disorders precede and predict the experience of hallucinations of various types, which would shed light on the causes of the development of hallucinations in some individuals with PD and not others. For example, our PD group without hallucinations included one participant who needed substantial contrast enhancement (69%); would this individual be more likely to eventually develop hallucinations than those who required less contrast?

In addition to attempting to understand the causes of hallucinations in PD, we are interested in potential interventions. Our findings indicated that when target contrast was enhanced, individuals with hallucinations, despite their lower contrast sensitivity baseline, were able to identify objects as accurately as those without hallucinations. An intriguing empirical question is whether perception‐based, non‐pharmacological interventions may change the type or reduce the frequency of occurrence of typically reported hallucinations and other misperceptions in PD. It may also be worth examining whether standard strategies for managing low vision in the general population (eg https://www.aota.org/About-Occupational-Therapy/Patients-Clients/Adults/LowVision.aspx), in older adults with concomitant cognitive impairment (eg ⁵⁸ ), and in disorders such as Alzheimer disease, ⁵⁹ , ⁶⁰ such as improving visibility by employing enhanced contrast in the home environment, allowing time for the eyes to dark adapt, and adjusting lighting to eliminate shadows and glare, may be beneficial to individuals with PD, or other disorders, who experience hallucinations.

Author Roles

(1) Research Project: A. Conception, B. Organization, C. Execution;

(2) Statistical Analysis: A. Design, B. Execution, C. Review and Critique;

(3) Manuscript: A. Writing of the first draft, B. Review and Critique.

M.D.S.: 1A,B,C; 2A,B; 3A

Z.A.M.: 1C; 2B; 3B

G.C.G.: 1A,B; 2C; 3B

S.N.: 2B,C; 3B

A.C.G.: 1A,B; 2C; 3B

Disclosures

Ethical Compliance Statement

The Boston University Institutional Review Board approved the study. Participants provided written informed consent according to the Declaration of Helsinki. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this work is consistent with those guidelines.

Funding Sources and Conflicts of Interest

This research was supported by a Research Supplement to Promote Diversity Award from the National Institute of Neurological Disorders and Stroke (NINDS) (R01 NS050446‐03S1), a Ruth L. Kirschstein National Research Service Award (F31NS074682), and a Clara Mayo Research Fellowship, Boston University, to MDS; a conference travel award from the Undergraduate Research Opportunities Program of Boston University to ZAM, and grants from NINDS (R01 NS067128, R01 NS050446) to ACG. The authors report no conflicts of interest.

Financial Disclosures for the Previous 12 Months

MDS was employed by the University of California, Los Angeles. She has received research funding including a Research Supplement to Promote Diversity in Health‐Related Research Award, National Institute on Aging (NIA), U01 AG052564‐01 Díaz‐Santos (PI Bookheimer); RCMAR/CHIME Pilot Grant NIA P30‐AG021684 (PI Mangione)/National Center for Translational Sciences UL1TR001881 (PI Dubinett). ZAM was employed by Duke University and Covance. He received a research grant, NIA F31AG060691. RDS was employed at Tewksbury Hospital. GCG was employed at Case Western Reserve University. He has received research funding on grant NSF PFI‐TT 2002721 (PI Lee, CWRU subaward PI Gilmore). SN was employed at Bridgewater State University. ACG was employed at Boston University. She has received funding on grants from the American Parkinson's Disease Association, NIH R01 AG063775 (PI Reinhart), and NIH R01 AG050595 (PIs Lyons, Kremen).

Supporting information

Appendix S1. Boston University Hallucinations and Unusual Perceptual Experiences Questionnaire (BU‐HUPE). BU‐HUPE was developed in our lab after consideration of the questionnaires and structured interviews available in the literature. The questionnaire included items from the University of Miami Parkinson's Disease Hallucinations Questionnaire (50), the North‐East Visual Hallucinations Interview (51), the Queen Square Visual Hallucination Inventory (52), and the Vision Questionnaire designed for PD (24, 53).