Interactive effects of locus coeruleus structure and catecholamine synthesis capacity on cognitive function

Abstract

Background: The locus coeruleus (LC) produces catecholamines (norepinephrine and dopamine) and is implicated in a broad range of cognitive functions including attention and executive function. Recent advancements in magnetic resonance imaging (MRI) approaches allow for the visualization and quantification of LC structure. Human research focused on the LC has since exploded given the LC's role in cognition and relevance to current models of psychopathology and neurodegenerative disease. However, it is unclear to what extent LC structure reflects underlying catecholamine function, and how LC structure and neurochemical function are collectively associated with cognitive performance.

Methods: A partial least squares correlation (PLSC) analysis was applied to 19 participants' LC structural MRI measures and catecholamine synthesis capacity measures assessed using [¹⁸F]Fluoro-m-tyrosine ([¹⁸F]FMT) positron emission tomography (PET).

Results: We found no direct association between LC-MRI and LC-[¹⁸F]FMT measures for rostral, middle, or caudal portions of the LC. We found significant associations between LC neuroimaging measures and neuropsychological performance that were driven by rostral and middle portions of the LC, which is in line with LC cortical projection patterns. Specifically, associations with executive function and processing speed arose from contributions of both LC structure and interactions between LC structure and catecholamine synthesis capacity.

Conclusion: These findings leave open the possibility that LC MRI and PET measures contribute unique information and suggest that their conjoint use may increase sensitivity to brain-behavior associations in small samples.

Keywords: catecholamine synthesis capacity; locus coeruleus; partial least squares correlation; positron emission tomography; structural magnetic resonance imaging.

Figure 1

The locus coeruleus and pontine reference masks in the group template space in the coronal (A), sagittal (B), and axial (C) view. The locus coeruleus (yellow arrows) and pontine reference (blue) regions of interest were manually segmented in the group brainstem template and then warped back to individual’s native space. The delineated locus coeruleus region was equally split into rostral (red), middle (green), and caudal (violet) parts in the rostrocaudal dimension.

Figure 2

Schematic representation of partial least squares correlation analysis (PLSC). The correlation matrix between the LC-MRI contrast ratios and LC MFT Ki_vis (A) or between the LC brain measures or neuropsychological performance (B) was subjected to singular vector decomposition. The association between the two latent variables after partialling out the chronological age and temporal day difference between cognitive and neuroimaging sessions represents the pattern of interindividual differences in LC-MRI contrast ratios that shares the largest amount of variance with interindividual differences in overall LC [¹⁸F]FMT Ki_vis (A) or that of interindividual differences in the LC brain measures that share the maximal covariance with interindividual differences in overall neuropsychological performance (B). MMSE, mini-mental state examination; WCS, Wisconsin card sorting task; WAIS-III, Wechsler adult intelligence scale III; WMS IV, Wechsler memory scale IV; CVLT, California verbal learning test.

Figure 3

Higher LC brain measures are associated with better cognitive functions. (A) Scatter plot depicting the association between latent LC brain measures and latent neuropsychological scores after partialling out chronological age and temporal day difference between cognitive and neuroimaging sessions. (B) Scatter plot indicating the distribution of the participants split into four subgroups by the mean of the LC brain measures. (C) Contribution of individual LC brain measures to a PLSC latent variable (bootstrap ratios >1.96 represented in red is considered reliable). LC, locus coeruleus; PLSC, partial least squares correlation.