Loss of functional connectivity is an early imaging marker in primary lateral sclerosis

Abstract

Objective: The clinical diagnosis of primary lateral sclerosis can only be made after upper motor neuron symptoms have progressed for several years without developing lower motor neuron signs. The goal of the study was to identify neuroimaging changes that occur early in primary lateral sclerosis, prior to clinical diagnosis.

Methods: MRI scans were obtained on 13 patients with adult-onset progressive spasticity for five years or less who were followed longitudinally to confirm a clinical diagnosis of primary lateral sclerosis. Resting state functional MRI, diffusion tensor imaging, and anatomical images were obtained. These "pre-PLS" patients were compared to 18 patients with longstanding, established primary lateral sclerosis and 28 controls.

Results: Pre-PLS patients had a marked reduction in seed-based resting-state motor network connectivity compared to the controls and patients with longstanding disease. White matter regions with reduced fractional anisotropy were similar in the two patient groups compared to the controls. Patients with longstanding disease had cortical thinning of the precentral gyrus. A slight thinning of the right precentral gyrus was detected in initial pre-PLS patients' scans. Follow-up scans in eight pre-PLS patients 1-2 years later showed increasing motor connectivity, thinning of the precentral gyrus, and no change in diffusion measures of the corticospinal tract or callosal motor region.

Conclusions: Loss of motor functional connectivity is an early imaging marker in primary lateral sclerosis. This differs from literature descriptions of amyotrophic lateral sclerosis, warranting further studies to test whether resting-state functional MRI can differentiate between amyotrophic lateral sclerosis and primary lateral sclerosis at early disease stages.

Trial registration: ClinicalTrials.gov NCT00015444 NCT01517087.

Figure 1.

Motor resting-state networks, showing differences in connectivity of the right motor cortex seed to the rest of the brain. (A) Red indicates brain regions with reduced connectivity in pre-PLS patients compared to heathy controls. (B) Red indicates brain regions in which established PLS patients have greater connectivity than in pre-PLS patients. (C) For eight pre-PLS patients with follow-up scans, regions in red had greater connectivity with the right motor cortex seed at the second timepoint (T2) compared to the first study (T1). Inset figure: 10-mm VOIs for generating the sensorimotor resting-state network were obtained from the voxel of peak activation in the left (MNI: −42, −20, 52; blue circle) and right (42, −16, 52; red circle) motor cortex from the right and left hand tapping task fMRI.

Figure 2.

Thinning of the motor cortex in PLS patients. (A) The mean thickness of the precentral gyrus was reduced in both hemispheres in established PLS patients (blue) compared to healthy controls (HC, black). Slight thinning of the right precentral cortex was seen in pre-PLS patients (red) at baseline. Asterisks indicate Bonferroni-corrected p < 0.05. Age, gender, and education were included as covariates. (B) In contrast, the postcentral gyrus had no significant thinning in PLS and Pre-PLS patients. (C, D) Longitudinal measures of cortical thickness of the precentral gyrus of 8 pre-PLS patients who had repeated scans. Cortical thickness is plotted against years since symptom onset. (C) Left precentral gyrus and (D) right precentral gyrus. Shaded area indicates the range of thickness for the healthy controls (mean ± 2 SDs).

Figure 3.

White matter changes in patients compared to controls. Regions of white matter with (A) reduced fractional anisotropy (red) in pre-PLS patients (top row) and established PLS patients (bottom row) compared to healthy controls. Regions of white matter with (B) increased mean diffusivity (blue) in pre-PLS patients (top row) and established PLS patients (bottom row) compared to healthy controls. There was no significant difference between pre-PLS and PLS patients (Tract-based spatial statistics, p < 0.05, FWE corrected.)