Posted June 21, 2023

Andre Irimia, Ph.D., University of Southern California

Traumatic brain injury (TBI) is a serious health problem for Veterans and civilians across the United States, with numerous studies showing it to be a major risk factor for late-life cognitive problems, including developing Alzheimer’s Disease (AD)/AD-related dementias (ADRD). Lack of diversity, reliance on self-report measures, and inaccurate/missing medical records have negatively contributed to the ongoing challenge of understanding just how TBIs lead to AD/ADRD. Physical markers that can be used to identify the disease early are also lacking. Symptoms of AD/ADRD can take many years to develop, so individuals are often unaware of risk factors that may increase their disease onset before it is too late. Therefore, it’s incredibly important that risk factors that speed up aging and AD/ADRD memory and thinking issues are identified as soon as possible.

Dr. Andrei Irimia at the University of Southern California received a fiscal year (FY) 2017 PRARP New Investigator Research Award that leveraged state-of-the-art machine learning to identify risk factors for cognitive decline and AD/ADRD progression using brain scans—functional magnetic resonance imaging (fMRI), diffusion tensor imaging (DTI), and resting-state activity of the default mode network (DMN)—in people as they age. These techniques allow researchers to examine how brain regions respond when performing various tasks at the same time as visualizing any brain damage. The main goal of this study was to use this information to identify which individuals with TBI have a higher risk of developing thinking and memory issues and AD/ADRD. Such early data are crucial to enable doctors to treat patients sooner with medicines, lifestyle changes, mental health measures, and health education to improve how individuals live with AD/ADRD.

In this PRARP-funded study, Dr. Irimia and his team first showed that geriatric mild TBI (mTBI; ages 57-79 years old) can lead to AD-like brain activity changes in the default mode network (DMN; Irimia, et al., 2020). Analysis of brain activity in the DMN is particularly important because it is a large-scale brain network that is particularly vulnerable to TBI because it contains long white matter tracts that get stretched or torn after injury. The DMNs of older individuals with mTBI exhibited distinct patterns of activity that are the same as AD patterns as early as ~ 6 months post-TBI. In both cases, there was a significant change in activity compared to healthy control subjects. These data suggest that older individuals are especially susceptible to the damaging effects of TBI on neural activity and AD-like disease progression.

Dr. Irimia recently demonstrated that this machine learning approach can identify differences between chronological age and MRI-derived brain age to systematically identify risk factors for decline post-TBI (Yin, et al., 2022). Machine learning uses using computers and artificial intelligence to analyze large amounts of data and “learn” from that data to predict results more accurately. This facilitates unbiased, sophisticated, computational analysis of numerous brain scans to detect even subtle differences between healthy and non-healthy aging brains (see Figure 1). A compilation of many brain scans, assessed by machine learning, produced detailed anatomic maps of how normal and TBI/AD-affected brains aged. Dr. Irimia’s team showed that, in individuals with MCI, these brain age estimates are significantly better than chronological age in capturing dementia symptom severity, functional disability, and executive function. This technology could be used to identify the earliest changes in individuals with TBI, thus enabling doctors to treat patients based on their brain-age-derived risk of developing MCI or AD/ADRD (Yin, et al., 2022).

Understanding TBI-related risk factors for the development of AD/ADRD are crucial for the significant aging population in the United States and abroad. The ability to use concrete biological data to quickly determine who is most at risk of developing thinking and memory problems is critical to ensuring that individuals are offered the best chance to find the most appropriate care approach (medical, lifestyle, mental health) to slow dementia. Further testing of this technology and approach is needed, but there is hope that this method can help change the standard of clinical care for TBI survivors.

References:
¹Irimia A, Maher AS, Chaudhari NN, et al. 2020. Acute cognitive deficits after traumatic brain injury predict Alzheimer's disease-like degradation of the human default mode network. Geroscience 42(5):1411-1429. doi: 10.1007/s11357-020-00245-6. Epub 2020 Aug 2. PMID: 32743786; PMCID: PMC7525415.

²Yin C, Imms P, Cheng M, et al. 2023. Anatomically interpretable deep learning of brain age captures domain-specific cognitive impairment. Proceedings of the National Academy of Sciences of the United States of America 120(2): e2214634120. doi: 10.1073/pnas.2214634120. Epub 2023 Jan 3. PMID: 36595679; PMCID: PMC9926270

Link:
Public and Technical Abstracts: TBI-Related Risk Factors for Alzheimer's Disease: Early Detection and Prognosis via Multimodal Brain Imaging and Connectomics

Top of Page