To analyze connectivity pattern of the default mode network (DMN) deactivation through the audio-motor transformation (AMT) in PD patients with different symptoms lateralisation.Parkinson's disease (PD) typically manifests with unilateral motor signs and subsequently evolve a number of motor and non-motor symptoms¹. Symptoms asymmetry is considered a crucial criterion for PD and may influence hemisphere-associated cognitive functions². PD patients may demonstrate cognitive decline at the earliest stage of disease³. DMN which is attributed to the resting brain activity regulation was shown to disrupt in a number of neuropsychiatric disorders and PD^3,4. DMN consists of posterior cingulate cortex (PCC), precuneus (pC), medial prefrontal cortex (MPFC), left and right inferior parietal lobuli (IPL_L, IPL_R), left and right posterior temporal lobe (TLP_L and TLP_R)⁴. It is still debated whether the lateralization of PD symptoms influences the way of future disease progression, especially concerning non-motor symptoms development and the role of DMN in this process^2,5. We hypothesized that AMT play important role in switching DMN to task-related deactivation state during voice-guided movements. We propose brain connectivity analysis during the voice-guided movements and AMT for DMN study in lateralized PD.Three groups of right-handed subjects (Gr1, Gr2, Gr3) were studied by fMRI with 1.5T SIGNA EXCITE (GE, USA). Gr1 consisted of 7 healthy subjects (4F, 51-83 y.o.). Gr2 consisted of 6 non-demented (MoCA 25-27) PD patients with left-side (non primary hand) motor symptoms (2F, 53-74 y.o.). Gr3 consisted of 5 non-demented (MoCA 25-30) PD patients with right-side (primary hand) motor symptoms (2F, 56-74 y.o.). For Gr2/Gr3 disease duration was 2-8 years, Hoen-Yahr scale 2.5-3. Simple reaction time in Gr1/Gr2/Gr3=302/352/328 ms. Simple finger tapping task was used for fMRI activation. Scanning session time was 3 min 24 s, 3 blocks of activation were acquired. Two estimation variables (EV) were generated. EV1 corresponded to the onset of voice command (AMT start/stop), EV2 corresponded to the movement execution solely. As the movement execution supposed predictable BOLD signal oscillation, estimated frequency for BOLD fluctuation for EV1 was ƒ₁=3.57x10^-2 Hz, and for EV2 was ƒ₂=1.78x10^-2 Hz. Single shot GE-EPI sequence was used for BOLD imaging (TR/TE=3000/71 ms, voxel size=4x4x5 mm). Anatomical images were acquired with FSPGR sequence (TR/TE=11.6/5.2 ms, TI=450 ms, voxel size=1x1x1.5 mm). FMRI data processing was carried out using GLM (FEAT) and ICA (MELODIC) based software from FSL (Oxford, GB). Standard FSL pre-processing was done. Single subject and group GLM and ICA analyses were done. ICA correlation analysis with GLM data was done using a built-in F-test. The frequency spectrum of BOLD signal fluctuations was analyzed for ICA components.

GLM and ICA based analyses of EV1 for Gr1, Gr2, Gr3 steadily revealed activation of neural network consisted of TLP_L and TLP_R, pC, supplementary motor area (SMA) and motor region of cingulate gyrus (Fig.1). The spectrum of the BOLD signal fluctuation for EV1 ICA component comprised one main frequency ν₁=3.57x10^-2 Hz (which is the same as ƒ₁). Same analyses for EV2 revealed activation of motor network comprising bilateral primary sensorymotor cortex, SMA, ventral premotor cortex and bilateral cerebellum. The spectrum of the BOLD signal fluctuation for EV2 ICA component comprised one main frequency ν₂=1.53x10^-2 Hz (which is the closest frequency to ƒ_2, do to the scanning temporat resolution). Although GLM didn’t showed DMN deactivation in Gr2, Gr3, ICA revealed DMN functioning exceptionally in EV2 anticorrelated manner (p<0.001) in Gr1, Gr2, Gr3 but not for EV1 or for the whole period of movement (main frequency in the ICA component spectrum for DMN was ν₃=1.53x10^-2 Hz=ƒ₂). The DMN topography was much more pronounced in Gr1, Gr2 and comprised all the described regions, while DMN structure for Gr3 consisted of pC, PCC, MPFC and the volume of TLP_L and TLP_R, pC nodes was much smaller than in Gr1, Gr2 (Fig.2). Functional connectivity through the region of posterior pC (MNI152: -2, -62, 46) was found for EV1 and for DMN deactivated during EV2 period.

The process of voice-guided movement execution and AMT in PD patients and age-matched subjects evokes activation of several neural networks. Auditory cortex, SMN and pC participate in AMT, which was previously shown⁶. In PD patients DMN is impaired⁷ and deactivates exceptionally at the period of motor activity. pC participates as a DMN ‘hub’⁸ and DMN deactivation ‘switching’ trigger during the AMT. Impaired AMT in voice-guided movement in PD might be another component of the freezing of the gate phenomenon. Primary-sided symptoms have more impact onto the DMN in PD.

PD symtoms laterality play important role in DMN functioning: primary sided symptoms prevalence influence DMN more then non-primary sided.