Cognitive decline in Parkinson's disease is associated with reduced complexity of EEG at baseline

Early detection of Parkinson's disease: Systematic analysis of the influence of the eyes on quantitative biomarkers in resting state electroencephalography

While resting state electroencephalography (EEG) provides relevant information on pathological changes in Parkinson's disease, most studies focus on the eyes-closed EEG biomarkers. Recent evidence has shown that both eyes-open EEG and reactivity to eyes-opening can also differentiate Parkinson's disease from healthy aging, but no consensus has been reached on a discriminatory capability benchmark. The aim of this study was to determine the resting-state EEG biomarkers suitable for real-time application that can differentiate Parkinson's patients from healthy subjects under both eyes closed and open. For this, we analysed and compared the quantitative EEG analyses of 13 early-stage cognitively normal Parkinson's patients with an age and sex-matched healthy group. We found that Parkinson's disease exhibited abnormal excessive theta activity in eyes-closed, which was reflected by a significantly higher relative theta power, a higher time percentage with a frequency peak in the theta band and a reduced alpha/theta ratio, while Parkinson's patients showed a significantly steeper non-oscillatory spectral slope activity than that of healthy subjects. We also found considerably less alpha and beta reactivity to eyes-opening in Parkinson's disease plus a significant moderate correlation between these EEG-biomarkers and the MDS-UPDRS score, used to assesses the clinical symptoms of Parkinson's Disease. Both EEG recordings with the eyes open and reactivity to eyes-opening provided additional information to the eyes-closed condition. We thus strongly recommend that both eyes open and closed be used in clinical practice recording protocols to promote EEG as a complementary non-invasive screening method for the early detection of Parkinson's disease, which would allow clinicians to design patient-oriented treatment and improve the patient's quality of life.

Automatic detection of Parkinson's disease from power spectral density of electroencephalography (EEG) signals using deep learning model

Parkinson's disease (PD) is characterized by slowed movements, speech disorders, an inability to control muscle movements, and tremors in the hands and feet. In the early stages of PD, the changes in these motor signs are very vague, so an objective and accurate diagnosis is difficult. The disease is complex, progressive, and very common. There are more than 10 million people worldwide suffering from PD. In this study, an EEG-based deep learning model was proposed for the automatic detection of PD to support experts. The EEG dataset comprises signals recorded by the University of Iowa from 14 PD patients and 14 healthy controls. First of all, the power spectral density values (PSDs) ​​of the frequencies between 1 and 49 Hz of the EEG signals were calculated separately using periodogram, welch, and multitaper spectral analysis methods. 49 feature vectors were extracted for each of the three different experiments. Then, the performances of support vector machine, random forest, k-nearest neighbor, and bidirectional long-short-term memory (BiLSTM) algorithms were compared using the PSDs feature vectors. After the comparison, the model integrating welch spectral analysis and the BiLSTM algorithm showed the highest performance as a result of the experiments. The deep learning model achieved satisfactory performance with 0.965 specificity, 0.994 sensitivity, 0.964 precision, 0.978 f1-score, 0.958 Matthews correlation coefficient, and 97.92% accuracy. The study is a promising attempt to detect PD from EEG signals and it also provides evidence that deep learning algorithms are more effective than machine learning algorithms for EEG signal analysis.

Spatio-Temporal Fractal Dimension Analysis from Resting State EEG Signals in Parkinson's Disease

Complexity analysis of electroencephalogram (EEG) signals has emerged as a valuable tool for characterizing Parkinson's disease (PD). Fractal dimension (FD) is a widely employed method for measuring the complexity of shapes with many applications in neurodegenerative disorders. Nevertheless, very little is known on the fractal characteristics of EEG in PD measured by FD. In this study we performed a spatio-temporal analysis of EEG in PD using FD in four dimensions (4DFD). We analyzed 42 resting-state EEG recordings comprising two groups: 27 PD patients without dementia and 15 healthy control subjects (HC). From the original resting-state EEG we derived the cortical activations defined by a source reconstruction at each time sample, generating point clouds in three dimensions. Then, a sliding window of one second (the fourth dimension) was used to compute the value of 4DFD by means of the box-counting algorithm. Our results showed a significantly higher value of 4DFD in the PD group (p < 0.001). Moreover, as a diagnostic classifier of PD, 4DFD obtained an area under curve value of 0.97 for a receiver operating characteristic curve analysis. These results suggest that 4DFD could be a promising method for characterizing the specific changes in the brain dynamics associated with PD.

Prediction of cognitive decline in Parkinson's disease (PD) patients with electroencephalography (EEG) connectivity characterized by time-between-phase-crossing (TBPC)

The aim of the study is to identify the dynamic change pattern of EEG to predict cognitive decline in patients with Parkinson's disease. Here we demonstrate that the quantification of synchrony-pattern changes across the scalp, measured using electroencephalography (EEG), offers an alternative approach of observing an individual's functional brain organization. This method, called "Time-Between-Phase-Crossing" (TBPC), is based on the same phenomenon as the phase-lag-index (PLI); it also considers intermittent changes in the signals of phase differences between pairs of EEG signals, but additionally analyzes dynamic connectivity changes. We used data from 75 non-demented Parkinson's disease patients and 72 healthy controls, who were followed over a period of 3 years. Statistics were calculated using connectome-based modeling (CPM) and receiver operating characteristic (ROC). We show that TBPC profiles, via the use of intermittent changes in signals of analytic phase differences of pairs of EEG signals, can be used to predict cognitive decline in Parkinson's disease (p < 0.05).

Machine Learning Identifies a Rat Model of Parkinson's Disease via Sleep-Wake Electroencephalogram

Alpha-synuclein induced degeneration of the midbrain substantia nigra pars compact (SNc) dopaminergic neurons causes Parkinson's disease (PD). Rodent studies demonstrate that nigrostriatal dopamine stimulates pallidal neurons which, via the topographical pallidocortical pathway, regulate cortical activity and functions. We hypothesize that nigrostriatal dopamine acting at the basal ganglia regulates cortical activity in sleep and wake state, and its depletion systemically alters electroencephalogram (EEG) across frequencies during sleep-wake state. Compared to control rats, 6-hydroxydopamine induced selective SNc lesions increased overall EEG power (positive synchronization) across 0.5-60 Hz during wake, NREM (non-rapid eye movement) sleep, and REM sleep. Application of machine learning (ML) to seven EEG features computed at a single or combined spectral bands during sleep-wake differentiated SNc lesions from controls at high accuracy. ML algorithms construct a model based on empirical data to make predictions on subsequent data. The accuracy of the predictive results indicate that nigrostriatal dopamine depletion increases global EEG spectral synchronization in wake, NREM sleep, and REM sleep. The EEG changes can be exploited by ML to identify SNc lesions at a high accuracy.

Deep learning reveals personalized spatial spectral abnormalities of high delta and low alpha bands in EEG of patients with early Parkinson's disease

Objective.Parkinson's disease (PD) is one of the most common neurodegenerative diseases, and early diagnosis is crucial to delay disease progression. The diagnosis of early PD has always been a difficult clinical problem due to the lack of reliable biomarkers. Electroencephalogram (EEG) is the most common clinical detection method, and studies have attempted to discover the EEG spectrum characteristics of early PD, but the reported conclusions are not uniform due to the heterogeneity of early PD patients. There is an urgent need for a more advanced algorithm to extract spectrum characteristics from EEG to satisfy the personalized requirements.Approach.The structured power spectral density with spatial distribution was used as the input of convolutional neural network (CNN). A visualization technique called gradient-weighted class activation mapping was used to extract the optimal frequency bands for identifying early PD. Based on the model visualization, we proposed a novel quantitative index of spectral characteristics, spatial-mapping relative power (SRP), to detect personalized abnormalities in the spatial spectral characteristics of EEG in early PD.Main results.We demonstrated the feasibility of applying CNN to identify the patients with early PD with an accuracy of 99.87% ± 0.03%. The models indicated the characteristic frequency bands (high-delta (3.5-4.5 Hz) and low-alpha (7.5-11 Hz) frequency bands) that are used to identify the early PD. The SRP of these two characteristic bands in early PD patients was significantly higher than that in the control group, and the abnormalities were consistent at the group and individual levels.Significance.This study provides a novel personalized detection algorithm based on deep learning to reveal the optimal frequency bands for identifying early PD and obtain the spatial frequency characteristics of early PD. The findings of this study will provide an effective reference for the auxiliary diagnosis of early PD in clinical practice.

Parkinsonism and dementia

The aim of the present review is to summarize literature data on dementia in parkinsonian disorders. Cognitive decline and the gradual development of dementia are considered to be key features in the majority of parkinsonian conditions. The burden of dementia in everyday life of parkinsonian patients and their caregivers is vast and can be even more challenging to handle than the motor component of the disease. Common pathogenetic mechanisms involve the aggregation and spreading of abnormal proteins like alpha-synuclein, tau or amyloid in cortical and subcortical regions with subsequent dysregulation of multiple neurotransmitter systems. The degree of cognitive deterioration in these disorders is variable and ranges from mild cognitive impairment to severe cognitive dysfunction. There is also variation in the number and type of affected cognitive domains which can involve either a single domain like executive or visuospatial function or multiple ones. Novel genetic, biological fluid or imaging biomarkers appear promising in facilitating the diagnosis and staging of dementia in parkinsonian conditions. A significant part of current research in Parkinson's disease and other parkinsonian syndromes is targeted towards the cognitive aspects of these disorders. Stabilization or amelioration of cognitive outcomes represents a primary endpoint in many ongoing clinical trials for novel disease modifying treatments in this field. This article is part of the Special Issue "Parkinsonism across the spectrum of movement disorders and beyond" edited by Joseph Jankovic, Daniel D. Truong and Matteo Bologna.

NREM Sleep EEG Characteristics Correlate to the Mild Cognitive Impairment in Patients with Parkinsonism

Early identification and diagnosis of mild cognitive impairment (MCI) in patients with parkinsonism (PDS) are critical. The aim of this study was to identify biomarkers of MCI in PDS using conventional electroencephalogram (EEG) power spectral analysis and detrended fluctuation analysis (DFA). In this retrospective study, patients with PDS who underwent an overnight polysomnography (PSG) study in our hospital from 2019 to 2020 were enrolled. Patients with PDS assessed by clinical examination and questionnaires were divided into two groups: the PDS with normal cognitive function (PDS-NC) group and the PDS with MCI (PDS-MCI) group. Sleep EEG signals were extracted and purified from the PSG and subjected to a conventional power spectral analysis, as well as detrended fluctuation analysis (DFA) during wakefulness, nonrapid eye movement (NREM) sleep, and rapid eye movement (REM) sleep. Forty patients with PDS were enrolled, including 25 with PDS-NC and 15 with PDS-MCI. Results revealed that compared with PDS-NC patients, patients with PDS-MCI had a reduced fast ratio ((alpha + beta)/(delta + theta)) and increased DFA during NREM sleep. DFA during NREM was diagnostic of PDS-MCI, with an area under the receiver operating characteristic curve of 0.753 (95% CI: 0.592-0.914) (p < 0.05). Mild cognitive dysfunction was positively correlated with NREM-DFA (r = 0.426, p = 0.007) and negatively correlated with an NREM-fast ratio (r = -0.524, p = 0.001). This suggested that altered EEG activity during NREM sleep is associated with MCI in patients with PDS. NREM sleep EEG characteristics of the power spectral analysis and DFA correlate to MCI. Slowing of EEG activity during NREM sleep may reflect contribution to the decline in NREM physiological function and is therefore a marker in patients with PDS-MCI.