Differences in Patterns of Atrophy in Alzheimer’s Disease and Psychosis
ABSTRACT
This research paper investigates brain gray and white matter atrophy patterns and their relationship to cognitive symptoms, such as poor visual processing, in three populations: individuals with Alzheimer’s disease (AD), individuals diagnosed with psychosis (PD), and age-matched healthy controls. Using data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI, N = 380) and the Human Connectome Project for Early Psychosis (HCP-EP, N = 242), volume, thickness, and atrophy rates were analyzed alongside visual processing performance. Age, structural patterns, and cognitive correlates were compared across these groups to identify shared and distinct features. Findings exhibited differences in atrophy, showing white matter degeneration in AD and gray matter atrophy in psychosis, both of which correlated with decreased visual functioning. This ultimately informs our understanding of differential atrophy pathology in neurodegenerative compared to psychiatric populations.
INTRODUCTION.
Differences in patterns across gray matter brain atrophy in Alzheimer’s disease (AD) and psychosis disease (PD) can provide crucial information regarding the mechanisms of impaired cognition, including deficits in visual processing, language, and memory. Both diseases involve brain atrophy, yet the regions impacted, and patterns of neurodegeneration differ across these populations. Atrophy, which also occurs in healthy aging [1], can occur due to water loss, synaptic degeneration, or tissue shrinkage. Most studies consider AD and PD separately, rather than directly comparing the two diseases. This lack of direct comparison limits our understanding of brain structural comparison between these populations. This gap is especially pronounced in work connecting regional gray matter loss to visual processing, an essential function for daily life. Previous literature has evaluated the impacts of regional atrophy on visual cognition and acuity within AD and PD separately, but little research has evaluated the cognitive impacts side by side. Comparative research has the potential to reveal distinct patterns of atrophy across the two populations and clarify how these structural changes contribute to cognitive deficits.
This study investigates patterns of hippocampal and ROI-based gray/white matter atrophy associated with cognitive symptoms, including visual processing deficits among individuals with AD, PD, and normal aging. By comparing atrophy across these groups, we sought to identify which structural changes are disorder-specific and which contribute to shared cognitive impairment. We hypothesized that AD would show widespread hippocampal atrophy, while PD will exhibit more localized, region-specific atrophy, yet both groups may display overlapping deficits in visual processing. Understanding these differences and similarities can illuminate how structural degeneration in the hippocampus translates into functional outcomes and may inform strategies for early detection and targeted intervention.
Background.
AD and PD affect a large portion of the population, and both are marked by structural brain changes, particularly by gray and white matter atrophy [2]. Both diseases have a strong negative impact on the brain, producing clinical symptoms such as reduced visual cognition, processing speed, and memory loss. The hippocampus, central to memory; temporal lobes, responsible for memory and language; and frontal lobes, critical for executive functions and working memory, seem most susceptible to gray matter atrophy and thinning in these conditions [3]. Longitudinal work has shown that AD patients lose roughly 4.66% of hippocampal volume annually, with atrophy accelerating around ages 60–70 and serving as a reliable marker beyond normal aging variability [1, 4]. Importantly, hippocampal volume has been shown to be a more sensitive biomarker than whole-brain atrophy for distinguishing AD, mild cognitive impairment (MCI), and healthy aging, predicting conversion from MCI to AD more effectively than global measures [4].
It is important to understand the differences in pathology between AD and psychosis, since hippocampal atrophy in AD is primarily driven by neuron loss rather than tau deposition alone, though tau and β-amyloid interactions contribute to degeneration [8]. AD shows variable atrophy patterns, hippocampal plus cortical, hippocampal only, and cortical only within the late stages of AD, while little to no atrophy can occur in early stages.
Unlike AD, where degeneration follows relatively predictable patterns, PD, particularly schizophrenia, presents with more region-specific and heterogeneous hippocampal abnormalities [9]. Late-onset PD is usually associated with frontotemporal involvement in those with the C9orf72 mutation [9]. Structural MRI studies reveal cortical thinning in visual areas such as the lateral occipital complex (LOC) and retinotopic cortex, with schizophrenia patients showing the most pronounced thinning compared with both healthy controls and bipolar patients [10]. These structural abnormalities align with functional impairments (e.g. visual masking tests), suggesting LOC and retinotopic deficits may underlie schizophrenia’s visual symptoms. [10]
Normal aging also involves hippocampal shrinkage, though typically at a slower and less disruptive rate [4]. A systematic study confirmed hippocampal atrophy as a baseline aging phenomenon, with methodological variability influencing reported rates, with minimal evidence for laterality or sex effects on the study [4]. Aging-related cognitive decline, however, differs qualitatively from AD. Comparative studies of dorsal stream visual processing show that while both groups display reduced heading and speed perception, AD patients have difficulty detecting consistent motion (higher coherence thresholds), whereas aging individuals struggle with perceiving changes in direction (higher direction thresholds), suggesting distinct mechanisms and areas of decline [11].
Together, these findings highlight hippocampal and visual cortical atrophy as shared but distinct phenomena across AD, PD, and normal aging. However, little work has examined how these structural differences translate into visual processing deficits, leaving a critical gap for integrative, interdisciplinary research.
MATERIALS AND METHODS.
Participants.
We used two large, publicly available datasets: the Alzheimer’s Disease Neuroimaging Initiative (ADNI) and the Human Connectome Project for Early Psychosis (HCP-EP).
Psychosis Cohort (HCP-EP): A total of 242 participants were included, consisting of 68 healthy controls (HC) and 174 individuals with psychosis (PD). The HCP-EP is a multi-site dataset that collects structural and functional MRI, cognitive, and clinical measures in early PD patients and matched controls, with standardized acquisition protocols to ensure comparability across sites.
Alzheimer’s Cohort (ADNI): Using the ADNI database, we analyzed data across a total of 257 participants, including 171 healthy controls as defined by ADNI diagnostic criteria, and 86 individuals diagnosed with AD. ADNI provides longitudinal neuroimaging, cognitive testing, genetic, and biomarker data across multiple diagnostic groups (HC, MCI, and AD), making it a widely used resource for studying structural brain changes and cognitive decline.
Participants missing cortical thickness, age, and diagnosis of severity levels were excluded from our analysis (N = 123).
Data Collection.
After securing data access, we curated a combined dataset. First, the full HCP-EP dataset was downloaded and organized into a working Excel sheet. Next, we reviewed ADNI’s data dictionaries to identify and match measures comparable to those present in HCP-EP, such as demographic variables and cognitive scores. These measures were extracted from ADNI and merged with HCP-EP into a master dataset containing participants from both cohorts.
In addition to core neuroimaging and cognitive data, we explicitly included age and sex for PD patients, as these variables are key covariates in both neurodegenerative and psychosis research. We were unable to access ADNI patient demographics.
Cognitive Measure: Clock Drawing Test.
As part of ADNI’s neuropsychological battery, participants completed the Clock Drawing Test (CDT) as part of the Montreal Cognitive Assessment (MoCA). Performance on the MoCA is scored on accuracy and organization. A higher score represents lower likelihood for cognitive impairment. The CDT is widely used as a measure of visuospatial processing, executive function, and attention, and impaired performance is considered an early indicator of cognitive decline in Alzheimer’s disease. In our study, CDT scores were used as a measure of visual processing ability, allowing us to test correlations between cognitive impairment and brain atrophy.
Analysis.
All analyses were performed using SPSS v29, a statistics software for advanced analysis, on the curated master dataset.
Demographic Comparisons.
We performed independent samples t-tests within each dataset to compare age and sex distributions between healthy controls and patient groups. For HCP-EP, comparisons were between HC and psychosis groups. For the ADNI data set, we were unable to obtain demographics due to HIPAA restrictions.
Group Differences in Cognition and Imaging Measures.
We ran additional independent t-tests to identify group differences in cognitive and structural measures, focusing on whether healthy vs. clinical populations showed significant divergence. To identify regions or measures of interest, we considered values with p < 0.05 as significant.
Correlational Analyses.
Within the ADNI dataset, we tested associations between CDT scores and atrophy measures using Pearson correlation coefficients. This allowed us to assess whether poorer visual/executive performance corresponded to greater volume loss or cortical thinning.
By combining these approaches, we were able to examine convergent patterns of atrophy and cognitive impairment in ADNI and structural/functional alterations in psychosis from HCP-EP, while maintaining consistent analytic pipelines across both disorders.
RESULTS.
Demographic Characteristics.
As shown in Table S1, demographic analysis revealed no significant group differences in age or handedness between healthy controls (HC) and individuals with psychosis (PD). However, participants in the psychosis cohort had significantly fewer years of education on average compared to controls (p = 0.001).
Cortical Thickness and Volume in Psychosis.
Demonstrated in Table S2 and figures S1-S4, significant reductions in cortical thickness and volume were observed in several regions among psychosis patients relative to controls. Specifically, psychosis participants exhibited decreased thickness in the right transverse temporal (p = 0.001), parahippocampus (p = 0.003), and fusiform regions (p = 0.008, p = 0.014), as well as in the right rostral middle frontal cortex (p = 0.03). The regions exhibit median cortical thickness that is consistently lower in the psychosis cohort compared to controls. Visual interpretation revealed a small number of extreme values in both groups, indicating heterogeneity in structural variation.
Cortical Thickness and Volume in Alzheimer’s Disease
Compared with healthy controls, participants with AD demonstrated widespread cortical and subcortical atrophy. Notably, significant reductions in volume were detected in the corpus callosum (anterior, mid-anterior, and posterior regions) and right accumbens area as shown in Table S3 and Figures S5-S8. Additionally, the right rostral anterior cingulate and left medial orbitofrontal cortices exhibited reduced cortical thickness in the AD cohort relative to controls. Overall, lower medians in the AD group compared to healthy controls were noted across all regions.
Cognitive Performance and Structural Correlations.
A boxplot of hippocampal volume compared to the MoCA CDT (Figure 1) demonstrated that participants with poorer test scores exhibited smaller left hippocampal volumes than high-performing individuals. Linear regression analysis (Figure 2) revealed a positive trend between left parahippocampal thickness and visual acuity; as parahippocamapal thickness increased, so did visual acuity. This suggests that structural preservation in the medial temporal lobe may contribute to better visuospatial performance.
Correlations Between Structural Measures and MOCA Subscores.
Across AD patients, significant correlations were observed between regional brain measures and MoCA CDT subscores. Table S4 shows that positive associations were found between total corpus callosum volumes and contour subscores (r = 0.156–0.214, p < 0.05), while negative correlations emerged between hippocampal thickness and contour scores (r ≈ –0.19, p < 0.03). Right parahippocampal thickness correlated negatively with MoCA Numbers subscores (r = –0.202, p = 0.002), indicating potential disruptions in visuospatial processing pathways.
DISCUSSION.
Atrophy Pattern Differences Between Alzheimer’s and Psychosis.
This study revealed consistent yet distinct patterns of cortical and subcortical atrophy between AD and psychosis cohorts compared to healthy controls. In AD, the most pronounced changes were observed in the corpus callosum, reflecting a predominately white matter pattern of degeneration.
In addition, hippocampal atrophy was evident, as shown in Figure 3, aligning with its role as an early and sensitive biomarker of AD. In contrast, PD patients showed atrophy primarily in gray matter regions, including the right transverse temporal lobe, right parahippocampal gyrus, left fusiform gyrus, and right rostral middle frontal lobe.
The divergence between these two conditions suggests different mechanisms underlying shared clinical symptoms such as impaired visual cognition, memory, and processing speed. White matter loss in AD likely impairs long-range communication between distributed brain networks, while gray matter thinning in psychosis reduces the integrity of local processing hubs [12]. This distinction emphasizes the need to consider tissue-specific patterns when evaluating neurodegenerative and psychiatric disorders.
Our results are consistent with expected findings based on previous literature. In AD, hippocampal degeneration is among the earliest and most severe hallmarks of disease progression. A reported annual 4.66% reduction in hippocampal volume in AD is shown to accelerate with age [1]. Similarly, medial temporal atrophy serves as a reliable biomarker for patients over the age of 85, though not exclusively for AD [5]. We confirm this result and extend it by identifying corpus callosum degeneration as an additional and significant marker of AD. This supports recent reports suggesting that white matter disconnection, rather than cortical atrophy alone, may play a critical role in AD-related cognitive decline.
By contrast, psychosis showed a markedly different anatomical signature. Whereas AD presented a relatively predictable trajectory of hippocampal and white matter loss, psychosis-related atrophy localized primarily in gray matter regions involved in perceptual and executive functions. This aligns with previous studies [10], that identified cortical thinning in visual areas such as the lateral occipital complex and retinotopic cortex in schizophrenia and associated late-onset psychosis with frontotemporal lobe atrophy [9]. Our data echo their findings, particularly in the fusiform and parahippocampal regions, suggesting that psychosis is characterized by local atrophy in cortical processing areas rather than global white matter.
Atrophy and Cognition: MOCA Clock Test.
One of the novel aspects of this study explores the correlation between structural atrophy and cognitive performance, as measured by the MoCA CDT in the AD cohort. Patients with reduced hippocampal and corpus callosum thicknesses/volumes performed significantly worse on visuospatial and executive aspects of the clock task. This finding directly bridges structural degeneration with measurable cognitive deficits, addressing a gap in academic literature. Our findings suggest that cortical thinning and white matter atrophy may translate into concrete impairments in visual and executive tasks.
Atrophy and Cognition: Visual Acuity in Psychosis Disease.
Additionally, the striking negative correlation between atrophy and visual acuity is seen within the PD population, suggesting that with increased atrophy rates, there is a decrease in visual acuity. In the parahippocampal regions, such findings connect degeneration with visual and occipital decline. Our findings suggest that parahippocampal thinning in PD could serve as an explanation for decreased visual acuity.
Limitations.
The limitations of this paper include age differences between the psychosis and Alzheimer’s Disease cohorts. The psychosis cohort consisted of individuals with mean ages of 24.29 years for HC, and 23.26 years for PD patients. Conversely, we assume that the AD cohort would have had average ages of much higher, creating a gap in data sets. However, due to the inaccessibility of demographics for the AD cohort, we do not have exact age data.
CONCLUSION.
This study shows that AD and PD, while sharing overlapping cognitive symptoms, are distinguished by distinct neuroanatomical atrophy patterns. AD patients exhibited pronounced white matter degeneration, whereas PD individuals showed gray matter thinning in localized areas. These findings support the hypothesis that structural patterns differ across disorders, with AD impairing long-range neural connection and communication, and psychosis disrupting the ability to process information. Correlations between hippocampal atrophy and poor performance on the MoCA Clock test further highlight the cognitive and functional impairments associated with structural decline, linking neuroimaging biomarkers to clinically accessible cognitive assessments.
The broader implication is that similar behavioral symptoms across psychiatric and neurodegenerative diseases may arise through distinct neurobiological pathways, supporting disorder-specific diagnosis. Future research should aim to compare white matter atrophy in AD and psychosis that may further reveal differences in connectivity disruptions. Expanding and integrating structural atrophy patterns with cognitive testing could improve early differential diagnosis and help tailor interventions.
ACKNOWLEDGMENTS.
Thank you to Faye McKenna from NYU for mentoring me through this process, and Lumiere Scholars Program for providing me with the opportunity to conduct this research.
SUPPORTING INFORMATION.
Supporting information available online includes:
Figure S1-S4: Boxplots showing psychosis region measure in comparison to healthy controls
Figure S5-S8: Boxplots showing Alzheimer’s region measure in comparison to healthy controls
Table S1: Demographic characteristics of healthy controls and psychosis group
Table S2: Regional cortical thickness and volume differences (PD)
Table S3: Regional cortical thickness and volume differences (AD)
Table S4: Correlations between brain measures and MOCA clock drawing sub scores (AD)
REFERENCES
[1] M. A. Fraser et al., A systematic review and meta-analysis of longitudinal hippocampal atrophy in healthy human ageing. NeuroImage 112, 364–374 (2015).
[2] F. Nabizadeh et al., Neuroimaging findings of psychosis in Alzheimer’s disease: A systematic review. Brain Behav. 15 (2025).
[3] D. C. Woodworth et al., Distinct patterns of gray matter atrophy in hippocampal sclerosis of aging and Alzheimer’s disease neuropathology. Alzheimers Dement. 16 (suppl. 2), e046089 (2020).
[4] W. J. P. Henneman et al., Hippocampal atrophy rates in Alzheimer disease: Added value over whole brain volume measures. Neurology 72, 999–1007 (2009).
[5] F. Barkhof et al., The significance of medial temporal lobe atrophy: A postmortem MRI study in the very old. Neurology 69, 1521–1527 (2007).
[6] Y. Chen et al., Pathology-specific patterns of cerebellar atrophy in neurodegenerative disorders. Brain (2025).
[7] R. de Flores et al., Contribution of mixed pathology to medial temporal lobe atrophy in Alzheimer’s disease. Alzheimer’s Dement. 16, 843–852 (2020).
[8] G. Halliday, Pathology and hippocampal atrophy in Alzheimer’s disease. Lancet Neurol. 16, 862–864 (2017).
[9] K. L. R. Socher et al., Multiple pathologies underlying a very late-onset schizophrenia-like psychosis. Alzheimer’s Dement. 19, e12345 (2023).
[10] E. A. Reavis et al., Cortical thickness of functionally defined visual areas in schizophrenia and bipolar disorder. Schizophr. Res. 179, 35–41 (2017).
[11] V. Kavcic, W. Vaughn, C. J. Duffy, Distinct visual motion processing impairments in aging and Alzheimer’s disease. Vision Res. 51, 386–395 (2011).
[12] Physiopedia, Grey and white matter. Physiopedia (2025); https://www.physio-pedia.com/Grey_and_White_Matter
Posted by buchanle on Tuesday, June 30, 2026 in May 2026.
Tags: ADNI (Alzheimer’s Disease Neuroimaging Initiative), Alzheimer’s Disease, HCP-EP (Human Connectome Project for Early Psychosis), Psychosis Disease