Background: The hippocampus is a critical brain structure for memory formation and other aspects of cognition. The hippocampus and the white matter tracts connecting it to other parts of the brain are known to lose volume and integrity with aging. For populations with prior compromised hippocampal integrity, such as those with autism spectrum disorder (ASD), it is less well known how the hippocampus and its connections will respond to aging. In children with ASD, there may be an initial period of enlarged hippocampi, after which there is a trajectory of faster decline in volume compared to neurotypicals (NT). We have previously identified reduced hippocampal volumes and fornix white matter integrity in middle-age and older adults with ASD compared to matched NT adults. However, freewater (FW) may be a more sensitive structural integrity measure of the hippocampal complex. FW is present in the brain as cerebrospinal fluid but also accumulates within the extracellular spaces indicative of reduced gray matter density and increased axon degeneration. FW shows promise as a more sensitive biomarker for Parkinson’s and Alzheimer’s disease. This study evaluated age-related hippocampal complex FW differences in adults with and without ASD across the adult lifespan. We hypothesized that adults with ASD would demonstrate a larger age association with increasing FW in the hippocampus and fornix, compared to NT adults, and that FW would be a more sensitive brain measure than traditional fractional anisotropy (FA).

Methods: The study consisted of 79 participants with ASD (59 male, 20 female; ages 18-70, mean=40.27 [±17] years) and 77 NT participants (46 male, 31 female; ages 18-71, mean=40.33 [±16] years). Hippocampal and fornix FW and FA values were generated from diffusion tensor images obtained along 32 directions using a b-value of 2500 s/mm2 in the axial direction with 3 mm slice resolution. These images were then processed for eddy current, distortion, b-vec and motion correction, skull stripped, and non-linear registered using Advanced Normalization Tools (ANTs) to the subject’s T1 image. FW and FA maps were calculated using custom written MatLab code and standard atlases containing the hippocampus and fornix were applied.

Results: The right hippocampus showed a significant diagnosis by age interaction (p=0.018), such that the increase in FW with age was greater for adults with ASD. The left hippocampus diagnosis by age interaction approached significance (p=0.055). Similarly, the right fornix showed a significant diagnosis by age interaction (p=0.044), with increases in FW with age as greater for adults with ASD, and the left fornix diagnosis by age interaction approached significance (p=0.053). FA values showed no significant diagnosis by age interactions.

Conclusion: In the hippocampus and fornix, the association between increasing FW and increasing age was more pronounced for adults with ASD than matched NT adults. This may mean that as adults with ASD age, these regions will degenerate faster than their NT peers, which could have implications for accelerated age-related memory decline. However, a notable limitation is the cross-sectional nature of the study. Our ongoing longitudinal study will inform a more definitive picture of brain aging with ASD.