Linking age changes in human neuronal microcircuits to impaired brain function and EEG biomarkers
Alexandre Guet-McCreight*1, Shreejoy Tripathy1,2,3,4, Etienne Sibille3,5,6, Etay Hay*1,2,3
1Krembil Centre for Neuroinformatics, Centre for Addiction and Mental Health, Toronto, Canada
2Department of Physiology, University of Toronto, Toronto, Canada
3Department of Psychiatry, University of Toronto, Toronto, Canada
4Institute of Medical Science, University of Toronto, Toronto, Canada
5Campbell Family Mental Health Research Institute, Centre for Addiction and Mental Health, Toronto, Canada
6Department of Pharmacology and Toxicology, University of Toronto, Toronto, Canada
*Email: Alexandre.Guet-McCreight@camh.ca and Etay.Hay@camh.ca
Introduction
Human brain aging involves a variety of cellular and synaptic changes, including loss of inhibitory cells [1], synaptic composition changes [2], and synaptic spine loss [3]. However, it remains poorly understood how these changes affect age-related impairments in human brain function. As our ability to study microcircuits in the living human brain is currently limited, detailed computational models of human brain microcircuits [4] offer a powerful tool to overcome limitations in linking microcircuit changes in aging to altered brain function and clinically-relevant brain signals [5].
Methods
We identified key human cellular and synaptic changes with age (loss of inhibitory cells, NMDA receptors, and spines), and integrated them into our detailed models of human cortical microcircuits [4]. Using these models of middle-age and older microcircuits, we simulated baseline and response spiking activity and resting-state electroencephalography (EEG) [6]. To characterize the effects of aging mechanisms on cortical function, we trained machine learning classifiers to detect signals based on readout of recurrent response vs baseline spiking. To characterize the effects on EEG, we analyzed EEG features using power spectral decomposition [7]. We then trained artificial neural networks [8] to estimate cellular aging from EEG biomarkers.
Results
Our detailed simulations of middle-age vs older human cortical microcircuits linked the altered aging mechanisms to reduced spike rates and impaired signal detection. Furthermore, we simulated EEG signals arising from the microcircuits and showed that the altered cellular mechanisms could account for key EEG power spectral biomarkers seen in human aging, including reduced aperiodic offset, exponent, and periodic peak center frequency. Each aging mechanism had unique effects on spiking and EEG. Using artificial neural networks, we showed that the aging mechanisms can be estimated accurately from EEG biomarkers.
Discussion
Our results overcome experimental limitations in linking cellular aging mechanisms with impaired cortical function and aging biomarkers in clinically-relevant brain signals. We demonstrate that cellular aging mechanisms have biomarkers in brain signals that can enable a more mechanistic stratification of older individuals with age-associated cognitive decline. Future studies can apply our in silico-trained artificial neural networks to estimate cellular aging for human subjects from their EEG data and assess the correspondence with cognitive impairments.
AGM and EH thank the Krembil Foundation for their generous funding support. AGM thanks the Canadian Institutes of Health Research - Institute of Aging for funding support.
[1] Chen et al (2023). 10.1016/j.neurobiolaging.2023.01.013 [2] Pegasiou et al (2020). 10.1093/cercor/bhaa052 [3] Petanjek et al (2011). 10.1073/pnas.1105108108 [4] Yao et al (2022). 10.1016/j.celrep.2021.110232 [5] Merkin et al (2023). 10.1016/j.neurobiolaging.2022.09.003 [6] Hagen et al (2018). 10.3389/fninf.2018.00092 [7] Donoghue et al (2020). 10.1038/s41593-020-00744-x [8] Abadi et al (2016). 10.48550/arXiv.1603.04467
