Feature Pattern Of Neural Oscillation And Network Connection In OSA Patients: A Sleep EEG Study

Objective Obstructive sleep apnea-hypopnea syndrome(OSA)is a sleep disorder disease,often accompanied by cardiovascular and cerebrovascular diseases which damages cognitive function.Polysomnogram(PSG)is a common method for clinical OSA detection,which can only qualitatively describe sleep disorders of OSA patients.Based on the PSG monitoring,the present study quantitatively assessed the abnormal oscillation in electroencephalography(EEGs)of OSA patients in different sleep stages and investigated the connection defect mechanism in sleeping EEGs network.This work was expected to provide support for the clinical diagnosis,treatment and pathogenesis study of OSA.Methods 1.ParticipantsOSA inclusion criteria was specified.OSA group is comprised of 10 OSA patients who met the criteria.Control group consists of 10 healthy subjects,selected and matched by age,sex and education from the OSA group.All patients underwent monitoring in the sleep respiratory therapy center at Tianjin Medical University General Hospital.2.PSG indicators PSG monitoring indicators were as follows: Apnea hypopnea index(AHI),the percentage of the time with Sa O2<90%(min)(% TST <90%),minimum oxygen saturation(min Sa O2),mean blood oxygen saturation(mean Sa O2),oxygen reduction index(ODI),micro arousal index(AI).3.Sleep EEGs data recording Sleep EEGs data was respectively recorded from each subject in 5 sleep periods: wake period(W phase),non-REM sleep stage 1(N1 stage),non-REM sleep stage 2(N2 stage),non-REM sleep stage 3(N3 stage)and rapid-eye movement sleep stage(R stage).10 segments of EEGs data in each stage were analyzed.Each segments lasts 10 s.4.EEGs data preprocessing Baseline drift,power frequency interference and artifacts(caused by EMG,ocular artifact,ECG)were removed from original recorded EEGs.5.Calculation and analysis of EEGs oscillation characteristic patterns in OSA Fast Fourier transform(FFT)method was used to calculate the power density of OSA EEGs in 5 sleep periods and control group and neural oscillations in the 5 periods were quantitatively assessed.The differences in power density between the two groups were compared to study the characteristic sleeping stages,frequency band and brain regions of OSA.6.Analysis of EEGs network connection characteristic patterns in OSADirected transfer function(DTF)of EEGs in the OSA and control group was calculated using the multivariable frequency domain Granger causality analysis method to quantitatively evaluate the causal network connection strength.The differences in network connection strength between the two groups were compared to study the characteristic sleeping stages,frequency band and brain regions of OSA.7.Statistical analysisAn independent sample t test was performed by using SPSS 22.0.Measurement data is expressed as mean ± standard error(Mean±SEM).Correlation analysis was done by using Pearson linear correlation.Results 1.PSG monitoring resultsPSG monitoring results showed that min Sa O2 and mean Sa O2 in the OSA group were lower than that in control group.% TST <90%,ODI index and AI index were higher than that in control group(P <0.05).2.EEGs oscillation characteristic patterns in OSA(1)In W stage,the power density in delta band in frontal and central regions of OSA is larger than that in Control(P <0.01);the power density in alpha band in parietal and occipital regions of OSA is lower than that in Control(P <0.05).(2)In N1 stage,the power density in delta band in frontal and central regions of OSA is larger than that in Control(P <0.01);(3)In N3 stage,the power density in delta and theta band in central region of OSA is larger than that in Control(P <0.05);3.EEGs network connection characteristic patterns in OSA(1)In W stage,the averaged DTF in gamma band in frontal region of OSA is larger than that in Control(P <0.05);the averaged DTF in gamma band in central region of OSA is lower than that in Control(P <0.05).(2)In N1 stage,the averaged DTF in beta band in central region of OSA is larger than that in Control(P <0.01).(3)In N3 stage,the averaged DTF in beta band in central region of OSA is larger than that in Control(P <0.05).Conclusion 1.PSG monitoring results suggest that the main symptoms of OSA are hypoxia and sleep disorders.2.The characteristic sleeping periods of OSA are W,N1 and N3 period.3.In W stage,the EEGs oscillations show enhancement in delta as well as weakening in alpha in OSA,indicating that the OSA patients are potentially sleepy in awake state.The EEGs show slow wave and alert function decline,which may be the compensatory mechanism of insufficiency of deep sleep;In N1 and N3 stages,the EEGs oscillations in delta and theta strengthen in OSA,indicating that the EEGs oscillations show slow wave in sleep and arousal response decline,which may be related with the sleep-awakening dysfunction.4.In W stage,the EEGs network connection in gamma band enhances in OSA,indicating that the brain functional connection of OSA patients shows a compensatory increase in awake state,in order to maintain attention,alert and other normal brain function.In N1 and N3 stages,the EEGs network connection in beta band strengthens in OSA,which may be related to control efforts of brain adjusting the respiratory muscle in apneic episodes,suggesting the compensatory mechanism for sensorimotor function dysfunction in brain with OSA.5.Abnormal EEGs oscillations and network connection were mainly found in the frontal and central regions in OSA,indicating that the characteristic changes in sleep EEGs of OSA may be related with the structural and functional damage of frontal and central regions.