Background Alcohol Use Disorder (AUD) is a risk factor for Alzheimer’s Disease (AD) and other dementias. Despite the emerging association between AUD and AD, there is a paucity of data available from mid‐life age ranges (e.g., 45 to 65 years) to assess AUD‐related risk for AD that might be present early in the disease trajectory. Method Functional connectome measures (graph‐theory) were calculated from resting state fMRI scans in 118 healthy controls (HC), 25 subjects with AUD (45‐61 years of age), and 51 subjects with amnestic mild cognitive impairment (aMCI) or AD. Following dimension reduction of 282 network nodes into 10 components using principal components analysis, predictive modeling with best subsets was used to isolate network components that predicted age in HC. Network components that significantly predicted age in HC were further examined to determine whether the AUD group exhibited premature brain aging or showed similar pathologic profiles as AD/aMCI. Result Predictive modeling with eigenvector centrality as a predictor of age yielded significant models (p’s < .0001). A network component including occipital and frontal pole nodes was a significant predictor in all models. In the frontal pole, AUD age trajectories showed a baseline shift relative to age‐matched HC subjects (49‐61 years of age), indicating premature brain aging. However, these age trends were not significantly different when Fisher z‐transformations of age trends were compared with a t‐test (p = .5). Nevertheless, the ANOVA revealed a significant main effect of group (AUD, HC, aMCI; p < .0001): average eigenvector centrality of the frontal pole in AUD was significantly lower than in HC (p = 0.0001) and indistinguishable from that of AD/aMCI subjects (p = 1.0), according to Bonferroni post‐hoc tests. Conclusion Premature brain aging in AUD may reflect vulnerability to AD‐like neurodegenerative processes that are manifest at the level of large‐scale network connectivity. Some data used in preparation of this abstract were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). As such, the investigators within the ADNI contributed to the design and implementation of ADNI and/or provided data but did not participate in analysis or writing of this report.

BackgroundPost‐Traumatic Stress Disorder (PTSD) and Alcohol Use Disorder (AUD) may be associated with premature brain aging which increases the risk for Alzheimer’s Disease and Related Dementias (ADRD). While each disorder alone has been studied as a risk factor for ADRD, few studies have examined premature brain aging in co‐occurring PTSD/AUD. It is important to know whether co‐occurring PTSD/AUD may amplify ADRD risk. Consequently, the present study tested the hypothesis that, relative to healthy controls (HC), PTSD/AUD would show more pronounced premature brain aging than AUD alone.MethodA functional connectome approach was used to develop a model of healthy brain aging. The graph‐theory measure eigenvector centrality was calculated from resting state fMRI scans in 60 mid‐life HC (HC‐mid; mean age=56.9), 69 older adult HC (HC‐older; mean age = 74.1), 36 AUD (mean age=54.4), and 30 PTSD/AUD participants (mean age=55.7). AUD and PTSD/AUD diagnoses were determined by a structured clinical interview. Connectome node features that were most predictive of age in HC‐mid and HC‐older were used in predictive modeling with bootstrap aggregation. These features were from multiple cortical areas and resting state networks. The healthy aging model was transferred to the AUD and PTSD/AUD groups to estimate brain age. Estimated ages in AUD and PTSD/AUD were compared to the predicted age in HC‐mid and HC‐older using generalized linear mixed modeling (GLMM).ResultPredictive modeling in HC yielded an aggregate model with 49.5% accuracy in predicting age (p < .0001), with some network features showing increases in eigenvector centrality with age and others showing decreases. After model transfer and correcting for age estimation bias, the GLMM yielded a main effect of diagnostic group (Wald c23=141.6, p < .0001). Predicted age for PTSD/AUD was significantly higher (M=61.0) than HC‐mid (M=57.2; p=.048) but predicted age for AUD (M=59.9) was not different from HC‐mid (p=.133). HC‐older predicted age (M=74.1) was different from all other groups (p < .0001).ConclusionIn this preliminary analysis, premature brain aging in PTSD/AUD was more pronounced than in AUD alone which may suggest greater vulnerability to AD‐like neurodegenerative processes that are manifest at the level of large‐scale network connectivity.

Background Alcohol Use Disorder (AUD) has previously been identified as a risk factor for Alzheimer’s disease and cognitive impairment. However, the mechanisms underlying the connection between AUD and Alzheimer’s disease remain unclear. Episodic memory is impaired in both AUD and mild cognitive impairment (MCI), and is associated with the transition from MCI to Alzheimer’s disease. This study used functional magnetic resonance imaging (fMRI) to compare activation during encoding of an episodic memory task in mid‐life AUD subjects to subjects with mild cognitive impairment (MCI), a prodromal stage of Alzheimer’s disease. Method Participants included 14 AUD subjects age 45 to 65 (M = 56 years), and 14 MCI subjects age 45 to 85 (M = 67 years). Healthy control subjects were matched to AUD and MCI groups by age and sex. Images were acquired on a 3T Siemens Prisma (3 mm3 voxels, TR = 1.1s, TE = 30 ms). During the fMRI task, participants viewed pictures of faces and objects and were instructed to remember these pictures for later. Activation maps were produced through general linear model analysis using cluster correction thresholding (Z>3.3, p = .05) in FSL (version 6.0). Results AUD and MCI groups showed reduced extent of BOLD activation during encoding versus baseline (Figure 1). Compared to matched controls, AUD and MCI subjects showed less medial temporal lobe (amygdala), brainstem, and thalamus activation. Additionally, AUD subjects did not exhibit the frontal activation that was present in MCI and matched controls. Conclusions Given the small sample sizes, caution should be taken in interpreting these results. Nonetheless, these preliminary findings suggest that mid‐life AUD subjects show similar patterns of reduced activation during episodic encoding to those with MCI. Reduced medial temporal lobe activation may represent vulnerability to Alzheimer’s disease, which is present in both MCI and AUD. Explanations for the differences in frontal activation might include compensatory mechanisms in MCI or executive dysfunction in AUD. Future analyses will aim to disentangle these effects by examining the relationship between BOLD signal activation and behavioral outcomes.

Background: Functional brain networks shift and adapt with task and environmental demands throughoutneural development and aging, which are both known to have gender specific profiles. The modularity ofthese networks has been shown to decrease with healthy aging but there is debate surrounding sexdifferences in aging and how dimorphisms relate to neurodegenerative pathologies such as Alzheimer's(AD). This study used resting state functional magnetic resonance imaging(fMRI) to examine sexdifferences in network modularity in mild cognitive impairment (MCI) and healthy control (HC) groups toexplore its potential as a clinical biomarker of a prodromal AD. Method: Participants included 90 MCI subjects age 55 to 77 (M = 68.19 years), and 127 HC subjects age45 to 84 (M = 66.53 years). Resting state fMRI scans were analyzed using standard preprocessing anddouble regression pipeline in FSL (version 6.0). BOLD time series were extracted from 280 brain regionsusing the Power atlas. Brain Connectivity Toolbox was used to compute the modularity graph theory metric.Generalized linear mixed modeling was used to examine the modularity metric Q as a function of sex andTrails-A score in HC and MCI.Result: A significant interaction between sex and Trails-A score was found (Wald c2=4.7, p=.031) withinthe HC group. Males showed a decrease in modularity with increased processing speed on Trails-A in linewith previous findings, whereas females showed no association. No significant interactions were found forMCI groups. Conclusion:These results confirm previous work correlating modularity with taskspecific connectivityand performance. Future analyses should focus on divergences between sexes in brain network modularityto inform differential diagnoses of AD.

BackgroundSubjective Cognitive Decline (SCD) may increase the risk of Alzheimer’s Disease and related dementias (ADRD). Brain network changes may occur at both the functional and structural levels in SCD. To better understand these network changes, diffusional kurtosis imaging (DKI) and functional connectome (FC) measures were examined in SCD and healthy controls (HC). We hypothesized that DKI and FC measures in the medial temporal lobe would be sensitive to differences between SCD and healthy controls (HC).MethodSCD classification was determined by clinician evaluation or Everyday Cognition average item score > 1.6, with objectively healthy cognitive performance (Montreal Cognitive Assessment score > 22). DKI and fMRI images were acquired on a Siemens PRISMA scanner in 33 SCD subjects and 33 age/sex matched HCs (mean age=69.1 yr). DSI Studio’s Automatic fiber tracking extracted the DKI measure mean kurtosis (MK) in 3 bilateral white matter bundles associated with ADRD: inferior longitudinal fasciculus (ILF), cingulum‐parahippocampus (CP), and uncinate fasciculus (UF). FC measure eigenvector centrality (Evc) was calculated in 19 medial temporal nodes using the Brain Connectivity Toolbox. Generalized Linear Models were conducted for each bundle to determine whether EVC in the left ventrolateral amygdala (LAMG) was predicted by MK, Diagnosis or the MK*Diagnosis interaction.ResultIn the left ILF, the effect of MK approached significance (p = .054) as did the MK*Diagnosis interaction (p = .059). Lower MK was associated with higher LAMG EVC in HC (r = ‐.40) but there was no association between MK and EVC for SCD. In the left UF, effect of MK was significant (p = .003), but no effect of Diagnosis and no interaction. Lower MK was associated with higher LAMG EVC (r = ‐.359) in both SCD and HC.ConclusionThis preliminary study showed that in HC, white matter integrity of the ILF was associated with LAMG functional connectivity, but this association was absent in SCD. While present results should be interpreted with caution due to small sample sizes, these findings indicate that healthy aging structural‐functional associations are disrupted in SCD, but future studies should determine whether this is a reliable marker of brain organization changes.