EEG biomarkers analysis in different cognitive impairment after stroke: an exploration study

 I hope you realize biomarkers DO ABSOLUTELY NOTHING FOR RECOVERY! And since you don't, everyone here is fired!

EEG biomarkers analysis in different cognitive impairment after stroke: an exploration study

Mengxue Xu^1†Yucheng Zhang^2†Yue Zhang³Xisong Liu^4*Kunqiang Qing^5,6*

• ¹Department of Neurology, Chongqing Public Healthy Medical Center, Chongqing, China
• ²Department of Mathematics, College of Natural Sciences, University of Texas at Austin, Austin, TX, United States
• ³Department of Psychology, School of Psychology, Shenzhen University, Shenzhen, China
• ⁴Intensive Care Unit, Chongqing Public Healthy Medical Center, Chongqing, China
• ⁵Automotive Software Innovation Center, Chongqing, China
• ⁶Research Group of Brain-Computer Interface, Brainup Institute of Science and Technology, Chongqing, China

Stroke is a cerebrovascular illness that brings about the demise of brain tissue. It is the third most prevalent cause of mortality worldwide and a significant contributor to physical impairment. Generally, stroke is triggered by blood clots obstructing the brain’s blood vessels, or when these vessels rupture. And, the cognitive impairment’s evaluation and detection after stroke is crucial research issue and significant project. Thus, the objective of this work is to explore an potential neuroimage tool and find their EEG biomarkers to evaluate and detect four cognitive impairment levels after stroke. In this study, power density spectrum (PSD), functional connectivity map, and one-way ANOVA methods were proposed to analyze the EEG biomarker differences, and the number of patient participants were thirty-two human including eight healthy control, mild, moderate, severe cognitive impairment levels, respectively. Finally, healthy control has significant PSD differences compared to mid, moderate and server cognitive impairment groups. And, the theta and alpha bands of severe cognitive impairment groups have presented consistent superior PSD power at the right frontal cortex, and the theta and beta bands of mild, moderated cognitive impairment (MoCI) groups have shown significant similar superior PSD power tendency at the parietal cortex. The significant gamma PSD power difference has presented at the left-frontal cortex in the mild cognitive impairment (MCI) groups, and severe cognitive impairment (SeCI) group has shown the significant PSD power at the gamma band of parietal cortex. At the point of functional connectivity map, the SeCI group appears to have stronger functional connectivity compared to the other groups. In conclusion, EEG biomarkers can be applied to classify different cognitive impairment groups after stroke. These findings provide a new approach for early detection and diagnosis of cognitive impairment after stroke and also for the development of new treatment options.

1 Introduction

Although the mortality rate of stroke has declined dramatically with the development of stroke research and advances in therapeutic techniques (1–4), disability from stroke continues to plague those who recover from stroke, with studies suggesting that as many as 30% of stroke patients have disability after recovery (5). This disability includes both post-stroke physical disability and post-stroke cognitive impairment (PSCI). PSCI is defined as any severity of cognitive impairment, regardless of cause, noted after an overt stroke (6, 7). This cognitive impairment involves multiple cognitive domains, with executive dysfunction being the primary and core symptom of PSCI (8–10). Depending on the severity of the cognitive impairment, PSCI can be categorized as mild, moderate, and severe. Mild PSCI consists of mild cognitive impairment that does not yet meet the diagnostic criteria for dementia. This may be manifested as memory loss, inattention, etc.; Moderate PSCI stage has more severe cognitive impairment, which may involve multiple cognitive domains, such as memory, language, and judgment. At this stage, the patient’s daily life may be significantly affected. Cognitive function is severely impaired when reaching severe PSCI, and the diagnostic criteria for dementia have been met. Patients may experience severe memory loss, language impairment, and behavioral abnormalities (11–14). Multiple studies have consistently demonstrated that individuals diagnosed with PSCI exhibit poorer rehabilitation outcomes related to physical function. Furthermore, they have a reduced likelihood of resuming a normal social life and exhibit a significantly higher mortality rate compared to those who do not have PSCI (15–18). Therefore, timely and accurate assessment of PSCI is extremely important, as it will help in the prevention and intervention of PSCI.

Currently there are two main types of assessments of the PSCI. One is questionnaire-based neuropsychological assessment, such as the Mini-mental State Examination (MMSE) and scales such as the Montreal Cognitive Assessment Scale (MoCA). However, this type of assessment is highly subjective, and the accuracy and reliability of the assessment results are more questionable. In contrast, biomarker-based assessment may be more objective, accurate and reliable. Specifically, EEG uses low-resistance electrodes placed on the human scalp to record oscillations generated by potential changes in the brain (19). EEG is a widely used non-invasive method for cognitive neurological research due to its high temporal resolution, ease of use, and low cost. A study that developed quantitative EEG (QEEG) to characterize EEG waves in post-stroke patients at risk of developing vascular dementia found that compared to normal subjects, patients with post-stroke with mild cognitive impairment had higher delta relative power, while alpha and beta relative power was lower in patients with post-stroke with mild cognitive impairment compared to normal subjects (20, 21). The study also examined the relationship between brain regions. The study also examined coherence between brain regions, with patients with post-stroke with mild cognitive impairment exhibiting lower interhemispheric and intrahemispheric coherences. Furthermore, event-related potentials (ERPs) were found to be lower in patients with mild cognitive impairment, while the relative power of alpha and beta was higher (20). ERPs derived from EEG have also been used to assess cognitive function in stroke patients. The P300 is sensitive in detecting subtle PSCI and can be used as an important marker for assessing PSCI, while P3 latency is an important marker of recovery from cognitive dysfunction after stroke (22, 23). From the perspective of the fNIRS, the PSCI group had lower intra-right and interhemispheric functional connectivity (FC) than healthy controls (HC). In the PSCI group, specific brain areas, such as the somatosensory cortex and prefrontal cortex, had considerably lower FC (24). Interestingly, neither acute ischemic stroke (AIS) patients nor the HC group showed prefrontal cortex (PFC) activation throughout the test in another study by researchers. (25). Arenth et al. found substantial changes in deoxy-Hb levels between aphasic and non-aphasic groups, but no differences between HC and non-aphasic stroke patients (26, 27). fMRI is another non-invasive neuromonitoring technique for monitoring blood oxygenation. fNIRS and fMRI are closely related, and studies have shown a significant correlation between the hemodynamic response measured by fNIRS and the blood oxygen level-dependent (BOLD) response obtained by fMRI (28). One resting-state fMRI study demonstrated that stroke affected both the lesioned and contralesional hemispheres through functional connectivity analysis of fMRI findings (29). While another study found that although both static and dynamic functional network connectivity varied in patients with PSCI, only the mean dwell time (MDT) metric in dynamic functional network analysis correlated with patients’ MMSE scores.

Despite the variety of biomarker-based monitoring methods, EEG is becoming increasingly popular in the research community and clinical practice by virtue of its low cost, high usability, and ease of setup (30–32). Due to these advantages, EEG may have higher generalizability compared to fNIRS and fMRI, and has great potential for the universal identification of PSCI. Considering that there is currently no unfied criteria for classifying the stages of PSCI, this research attempts to divide PSCI into four stages from NC to server cognitive impairment, and tries to find the differences in EEG biomarker on these four levels of impairment, which on one hand, improves the framework of dividing the symptomatic stages of PSCI, and on the other hand, helps clinical workers to assess the stage of impairment of patients with PSCI, and facilitates individualized interventions and preventions for them.