White matter microstructure predicts cognitive training-induced improvements in attention and executive functioning in schizophrenia

Abstract

We examined the relationship between white matter microstructure in schizophrenia using diffusion tensor imaging (DTI) and cognitive improvements induced by 70h (~16weeks) of cognitive training. We measured anatomical connectivity in 48 patients with schizophrenia (SZ) and 28 healthy control participants (HC) at baseline, and then examined the relationship between anatomical connectivity at baseline and training-induced cognitive gains in 30 SZ who performed diffusion imaging after completing 70h of training. Compared with healthy control participants, individuals with schizophrenia showed reduced white matter integrity at baseline, as indexed by fractional anisotropy metrics, in bilateral posterior corona radiata, bilateral retrolenticular internal capsules, bilateral posterior thalamic radiation, left anterior corona radiata, left superior longitudinal fasciculus, left sagittal stratum, right cerebral peduncle and the genu and splenium of the corpus callosum. After training, schizophrenia participants showed significant gains in attention/vigilance, speed of processing, verbal learning, visual learning and executive functioning. White matter integrity within the right fronto-occipital fasciculus predicted training-induced improvements in attention/vigilance, while white matter integrity within the right corticospinal tract and bilateral medial lemnisci predicted cognitive training-induced improvements in executive functioning, areas that did not show white matter tract deficits at baseline. These findings suggest that preserved white matter integrity connecting long-range prefrontal-thalamic-sensorimotor areas may be an important determinant for training-induced neurocognitive plasticity.

Keywords: Attention; Cognitive training; Diffusion tensor imaging; Executive functioning; Prefrontal cortex; Schizophrenia.

Figure 1

Tracts showing significant reduction in FA in SZ patients (N = 48) compared with HC participants (N = 28), (regional reductions illustrated in blue voxels, TFCE, p_FWE < 0.05). Results are overlaid on the study-specific whole-group mean FA image and the whole-group mean FA skeleton.

Figure 2

Front-occipital, corticospinal and medial lemniscus tracts illustrated in red, reveal tracts in which FA was not significantly different in SZ (N=48) vs. HC (N=28) participants, indicating relatively preserved white matter integrity within these tracts in SZ patients. Blue areas show tracts with reduced FA values in SZ vs. HC participants. Greater FA at baseline in right fronto-occipital tract is significantly associated with larger cognitive training-induced improvements in attention, and greater FA in right cortico-spinal tract and bilateral medial lemnisci is significantly associated with greater cognitive training-induced improvements in executive functioning in SZ patients who completed 70 hours of training (N=30). Tracts were identified using the Johns Hopkins University Probabilistic Tractography and White Matter Labels Atlases (distributed with FSL).