P 38. Mild cognitive impairment in patients with Parkinson's disease is characterized by a strong coupling of EEG signal complexity and band power

Abstract

Introduction. Mild cognitive impairment (MCI) in Parkinson“s disease (PD) has been recognized as a transitional state between normal cognition (PD-NC) and overt dementia (PD-D) and is associated with an increased risk of conversion toward PD-D. In a cross-sectional setting the combination of spectral band power and signal complexity for use as a diagnostic biomarker for assessing MCI in early stages of PD is evaluated. Patients & Methods. 30 PD-NC individuals (13 females, medians: age 67.5 yrs., edu. 14 yrs., UPDRS 16, LED 507.5, disease duration 33 mos.) were compared with 9 individuals diagnosed with PD-MCI matched for age, sex and education (4 females, medians: age 66.0 yrs., edu 16 yrs., UPDRS 15, LED 770.0, disease duration 40 mos.). High-density EEG recordings were obtained during eyes open (EO) and eyes closed (EC) condition. Signal complexity, quantified based on Tsallis entropy (TE) [1], as well as relative band power (rBP) were evaluated for the delta, theta, alpha, beta and gamma band. For all bands and conditions, the Pearson correlation between TE and rBP was calculated for each individual. The resulting correlation scores were then compared across groups according to an individuals“ diagnosis. Results. TE and rBP analyzed independently, showed no significant differences between PD-NC and PD-MCI. However, in combination, significant differences were observed: Compared to PD-NC, most patients with PD-MCI showed significantly higher correlation of TE and rBP in both the alpha band (8-13Hz) in EO condition as well as the beta band (13-30Hz) in EC condition. The corrected p-values in both cases were p<0.05. Furthermore, in case of PD-MCI, the median correlation was negative: -0.33 (alpha band, EO) and -0.50 (beta band, EC). For PD-NC individuals, median correlation was positive: 0.11 (alpha band, EO) and 0.09 (beta band, EC). Other bands, in either condition, showed no significant group differences in correlation between TE and Rbp. Conclusion/Discussion. In conclusion, signal complexity and relative band power provide, in combination, a candidate diagnostic biomarker for MCI in PD. Higher correlation in case of PD-MCI indicates a reduced ability of the cortex to modify signal complexity independently of relative band power. The results apply to groups, and may help to define study populations for clinical trials, but cannot be applied in their present form for individual treatment decisions or counselling. Reference [1] Sneddon R. The Tsallis entropy of natural information. Physica A Statistical and Theoretical Physics 2007; 386: 101–118. Fig. 1.