Neurotransmitter systems explain lifespan changes of human resting-state brain activity

1248 

Abstract Submission 

Leon Lotter^1,2,3, Jan Kasper^1,2, Simon Eickhoff^1,2, Juergen Dukart^4,2

¹INM-7, Research Centre Jülich, Jülich, NRW, Germany, ²Institute of Systems Neuroscience, Heinrich Heine University, Düsseldorf, NRW, Germany, ³Max Planck School of Cognition, Leipzig, Saxony, Germany, ⁴INM-7, Forschungszentrum Jülich, Jülich, NRW, Germany

Leon Lotter  
INM-7, Research Centre Jülich|Institute of Systems Neuroscience, Heinrich Heine University|Max Planck School of Cognition
Jülich, NRW, Germany|Düsseldorf, NRW, Germany|Leipzig, Saxony, Germany

Jan Kasper  
INM-7, Research Centre Jülich|Institute of Systems Neuroscience, Heinrich Heine University
Jülich, NRW, Germany|Düsseldorf, NRW, Germany

Simon Eickhoff  
INM-7, Research Centre Jülich|Institute of Systems Neuroscience, Heinrich Heine University
Jülich, NRW, Germany|Düsseldorf, NRW, Germany

Juergen Dukart  
INM-7, Forschungszentrum Jülich|Institute of Systems Neuroscience, Heinrich Heine University
Jülich, NRW, Germany|Düsseldorf, NRW, Germany

Human brain activity, as measured using resting-state functional (rsf)MRI, changes during neurodevelopment and aging (Ferreira & Busatto, 2013; Hu et al., 2014; Uddin et al., 2010). Despite solid evidence for age-related alterations, their interpretation often remains unspecific, due to a lack of non-invasive methods that provide insights into the underlying neurobiological processes. We approach this problem using a framework that was previously successfully applied to identify potential neurobiological mechanisms contributing to structural brain development (Lotter et al., 2023): We first establish regional developmental trajectories of an rsfMRI measure of brain activity – the fractional Amplitude of Low Frequency Fluctuations (fALFF) (Zou et al., 2008) – in a lifespan sample of 2444 individuals. Using these modeled trajectories, we construct "representative" fALFF change brain maps across the lifespan. We hypothesize that fALFF spatial change patterns (i.e., stronger vs. weaker change in one vs. another brain region during a given age span) reflect developmental changes in specific neurobiological systems (Dukart et al., 2021; Lotter et al., 2023). For this, we compute spatial colocalization analyzes testing the extent to which lifespan fALFF changes are explained by distributions of specific neurotransmitter systems and brain metabolism.

We employed the lifespan Human Connectome Project dataset (HCP-D: n = 642, 5–22 years; HCP-YA: n = 1093, 22–37 years, HCP-A: n = 709, 36–90 years). HCP-preprocessed MRI data was cleaned from white matter and CSF signals, parcellated into 416 cortical and subcortical regions (Huang et al., 2022), and fALFF was calculated on each region's time series (0.01–0.1 Hz). We used "normative" warped Bayesian linear regression models to model fALFF per brain region across age, accounting for effects of sex, study site, and motion (Rutherford et al., 2021). Predictions from these models were used to construct fALFF change maps across 5-year time windows from 5 to 90 years. As potential biological correlates of fALFF change patterns, we gathered and parcellated a collection of group-average PET maps (Markello et al., 2022). To account for their intercorrelation, we applied hierarchical clustering to group them according to their spatial correlation. Following this hierarchy, we used linear regression models to examine if fALFF change patterns are explained by specific PET-derived neurotransmitter and metabolism maps (outcome: fALFF change, predictor(s): PET maps in a cluster, observations: parcel-wise values, controlled for partial volume effects). Statistical significance was established using spatial null maps and FDR correction (Lotter et al., 2023; Lotter & Dukart, 2022).

Developmental models explain up to 56% of variance in fALFF data and revealed a general trend towards lower brain activity with higher age in mostly transmodal cortical brain areas (examples in Fig. 1A). PET maps cluster broadly according to receptors with high density in subcortical areas (high-striatal vs. high-thalamic) and those concentrate in cortical areas (brain metabolism vs. "general": serotonergic, glutamatergic, GABAergic transmitter systems; Fig. 1B). Modeled fALFF change patterns are explained to large extents by both main clusters (multivariate R2 up to 49%; Fig. 2: top), with the strongest contribution by major cortical transmitter systems (multivariate R2 up to 42%; center) and an emphasis on glutamatergic and GABAergic receptors (univariate R2 up to 29%; bottom).

Development of human brain activity measured with rsfMRI follows the distributions of specific neurotransmitter systems. Our approach not only identifies neurobiological mechanisms associated with human brain-functional development, but also constitutes the foundation for future biologically grounded biomarkers based on individual functional MRI.

Early life, Adolescence, Aging ¹

Multivariate Approaches

Normal Development

Transmitter Systems ²

Pharmacology and Neurotransmission

Development

FUNCTIONAL MRI

GABA

Glutamate

Multivariate

Neurotransmitter

Other - Resting-state fMRI; Spatial Colocalization; Spatial Correlation