Pattern of cortical thinning in logopenic progressive aphasia patients in Thailand

Abstract

Background: Logopenic progressive aphasia (LPA) is an uncommon neurodegenerative disorder primarily characterized by word-finding difficulties and sentence repetition impairment. Prominent cortical atrophy around left temporo-parietal junction (TPJ) is a classical imaging feature of LPA. This study investigated cortical thinning pattern in clinically diagnosed LPA patients using non-demented subjects as a control group. We also aimed to explore whether there was prominent thinning of other cortical area additional to the well-recognized left TPJ.

Methods: Thicknesses of all cortical regions were measured from brain magnetic resonance images using an automated command on Freesurfer software. Cortical thickness of the LPA and control groups were compared by two methods: 1) using a general linear model (GLM) in SPSS software; and 2) using a vertex-by-vertex GLM, performed with Freesurfer's QDEC interface.

Results: Besides the well-recognized left TPJ, cortical regions that were significantly thinner in the LPA group by both comparison methods included left caudal middle frontal gyrus (CMFG) (p = 0.006 by SPSS, p = 0.0003 by QDEC), left rostral middle frontal gyrus (p = 0.001 by SPSS, p = 0.0001 by QDEC), left parahippocampal gyrus (p = 0.008 by SPSS, p = 0.005 by QDEC) and right CMFG (p = 0.005 by SPSS, p = 0.0001 by QDEC).

Conclusions: Our results demonstrated that thinning of middle frontal gyri may be an additional feature in clinically diagnosed LPA patients. Involvement of left parahippocampal gyrus may reflect the underlying neuropathology of Alzheimer's disease in majority of the LPA patients.

Keywords: Alzheimer’s disease; Aphasia, primary progressive; Cortical thickness; Freesurfer; Logopenic progressive aphasia; Magnetic resonance imaging.

Fig. 1

Cortical thickness of each cortical region in LPA and control groups. *Using GLM controlling for age, sex and total brain volume in SPSS software, LPA group was shown to have significant cortical thinning on the left cerebral hemisphere in the banks of the superior temporal sulcus (BSSTS) (p = 0.001), caudal middle frontal gyrus (CMFG) (p = 0.006), entorhinal cortex (EC) (p = 0.001), lateral occipital gyrus (LOG) (p = 0.005), middle temporal gyrus (MTG) (p = 0.002), parahippocampal gyrus (PHG) (p = 0.008), rostral middle frontal gyrus (RMFG) (p = 0.001) and supramarginal gyrus (SG) (p = 0.006), while on the right hemisphere in the BSSTS (p = 0.006), CMFG (p = 0.005), EC (p = 0.002), MTG (p = 0.005), PHG (p = 0.002) and temporal pole (TP) (p = 0.005)

Fig. 2

Comparing cortical thickness between groups using vertex-by-vertex GLM analysis in Freesurfer’s QDEC interface. Clusters showing significant cortical thinning in the LPA than in control groups are demonstrated on the cortical surface templates. On the left hemisphere, these clusters locate in the middle temporal gyrus (MTG) (area 1117.81 mm², p = 0.0001), supramarginal gyrus (SG) (area 932.98 mm², p = 0.0001), rostral middle frontal gyrus (RMFG) (area 791.38 mm², p = 0.0001), caudal middle frontal gyrus (CMFG) (area 435.78 mm², p = 0.0003), superior frontal gyrus (SFG) (area 334.68 mm², p = 0.003), parahippocampal gyrus (PHG) (area 296.81 mm², p = 0.005), temporal pole (TP) (area 290.13 mm², p = 0.005), and inferior temporal gyrus (ITG) (area 264.06 mm², p = 0.01). On the right hemisphere, the cluster locates in the CMFG (area 503.43 mm², p = 0.0001). There was no cluster showing significantly thinner cortex in the control than in LPA groups