Decreased and increased cerebral regional homogeneity in early Parkinson's disease

Highlights

• We examine altered regional homogeneity of the brain in Parkinson's disease by fMRI.

• Homogeneity in the right primary sensory cortex was decreased in Parkinson.

• Homogeneity in the left angular gyrus was increased in Parkinson.

• Homogeneity in the right primary sensory cortex was decreased in early Parkinson.

• Homogeneity in the right primary sensory cortex was associated with disease duration.

Abstract

Regional homogeneity (ReHo), a processed data from resting-state functional magnetic resonance imaging, provides information about spontaneous brain activity in focal areas. Altered ReHo in brain regions has been reported in Parkinson's disease (PD). We compared ReHo of 22 patients with PD in their off-medication state to 25 healthy controls. We observed decreased ReHo in the right primary sensory cortex, the right primary motor cortex and the right middle frontal gyrus in PD patients. Conversely, ReHo was increased in the left inferior parietal lobule, the angular gyrus, the supramarginal gyrus, the middle occipital gyrus and the parahippocampal gyrus. In the right primary sensory cortex, ReHo showed a positive association with disease duration which proposed the low level of ReHo in the early phase of PD. ReHo was decreased in the seven de novo PD patients with disease duration less than 1 year as compared to the control which corresponded to the prediction. In both the off-medication and de novo PD patients, ReHo decreased in the right primary sensory cortex and increased in the angular gyrus as compared to the control. Potential regression of ReHo extrapolated backwards to PD-onset may provide a clue for ‘premotor diagnosis’.

Introduction

Parkinson's disease (PD) is a progressive neurodegenerative disease characterized by distinct motor symptoms such as rigidity, tremor and slowness of movement. These motor symptoms are not manifested until the nigrostriatal dopaminergic circuit is damaged severely, which results in late diagnosis. In the advanced stage of PD, no medication can rehabilitate the neural degeneration; neuroprotective therapies are also not satisfactory. Detection techniques of early PD are desired for this reason.

Single photon emission computerized tomography and positron emission tomography, the most common imaging techniques of PD (Niethammer et al., 2012), are efficient in distinguishing PD from other neural disorders and are potentially powerful early diagnosis techniques of PD (Cummings et al., 2011). However, the invasiveness caused by the utilization of radioactive tracers is a limitation to those techniques. A non-invasive tool such as functional magnetic resonance imaging (fMRI) can be a promising alternative detection technique of PD.

Among the various application of fMRI, regional homogeneity (ReHo), a kind of processed fMRI signal, has been utilized in automatic discrimination of early PD (Long et al., 2012). Generally, ReHo analyses were applied to the resting-state fMRI (RS-fMRI) data which records the blood oxygen level-dependent (BOLD) signals while participants lie still inside an MRI scanner. The RS-fMRI is thought to reflect more spontaneous brain status than task-related fMRI (TR-fMRI) which triggers brain activation artificially (Greicius et al., 2003, Fox and Raichle, 2007, Fox and Greicius, 2010).

Brain navigation by ReHo is grounded on a hypothesis: a region with a high ReHo can be a topical functional unit (Zang et al., 2004). ReHo indicates the level of intraregional similarity of small brain regions and is acquirable by calculating the Kendall's coefficient of concordance (KCC; Kendall and Gibbons, 1990) with BOLD signals. BOLD signals have been suggested to reflect neural activity (Logothetis et al., 2001), especially in the low-frequency band (<0.8 Hz, Biswal et al., 1995). Coherence of BOLD fluctuations in a region can be indicative of functional similarity of the region in regards to neural activation.

Further studies exhibiting the association between ReHo and cerebral hemodynamics supports this hypothesis (Li et al., 2012b, Zou et al., 2009). Reliability of the ReHo approach is underpinned by the test–retest confidence in almost all gray matter (Li et al., 2012a). ReHo has been found to have associations with gray matter concentration (Lv et al., 2011), voluntary movement (Zang et al., 2004), normal aging (Wu et al., 2007) and hearing (Li et al., 2012). Disorder-dependent ReHo changes have been investigated in Alzheimer's disease (He et al., 2007), schizophrenia (Liu et al., 2006), attention deficit hyperactivity disorder (Cao et al., 2006, Zhu et al., 2008) and multiple system atrophy (You et al., 2011).

PD-dependent ReHo changes have also been observed (Wu et al., 2009a). In the study, ReHo of PD patients decreased in the putamen, thalamus and supplementary motor area while exhibiting an increase in the cerebellum, primary sensorimotor cortex (SMC) and premotor area. These observations were made with a temporary cessation of medication (off-medication state). Interestingly, these ReHo alterations were normalized by instant administration of l-dopa in the study.

With this modulatory function of l-dopa on ReHo, we recalled TR-fMRI studies that presented controversial BOLD activations between de novo PD patients (early and drug-naïve PD patients) and off-medication PD patients. According to the basal ganglia – thalamocortical circuit model of PD (Alexander et al., 1990), the primary motor cortex (M1), a part of SMC, will show decreased activity due to a reduction in excitatory thalamic outflow (Albin et al., 1989). This hypothesis was supported by the regional cerebral blood flow measurement (Rascol et al., 1992). However, contradictory hyperactivations of the M1 were reported in TR-fMRI studies which involved motor-task, (Sabatini et al., 2000, Haslinger et al., 2001, Wu and Hallet, 2005, Eckert et al., 2006, Moraschi et al., 2010).

The hyperactivation was refuted by the studies performed in de novo PD patients, where M1 showed hypoactivation (Buhmann et al., 2003, Tessa et al., 2010, Tessa et al., 2012). The hyperactivation was assumed to reflect reorganization of the M1 in virtue of compensatory mechanism caused by l-dopa treatment (Haslinger et al., 2001, Eckert et al., 2006). If l-dopa normalizes the disturbed ReHo, even though it affects temporarily (Wu et al., 2009a), long-term medication may also reorganize ReHo as it has done in the TR-fMRI.

The aim of this study was to examine whether ReHo alteration observed in off-medication PD patients is also observable in de novo PD patients. We investigated the correlation between ReHo and clinical features to find a clue for early detection of PD. To achieve these goals, we observed ReHo alteration in the off-medication PD group compared to the healthy group and defined the altered regions as regions of interest (ROIs). The ReHo values of the ROIs in the de novo PD patients were measured and compared with those of the off-medication PD and the control group.