Coma is a common clinical symptom among neurocritical patients, characterized by a state of unconsciousness and an incapacity to consciously and purposefully react to stimuli from the external environment, often indicating a significant decline in cerebral functionality (Huff and Tadi, 2023). Coma is classified into two types: traumatic and non-traumatic coma. The prognostic assessment of coma patients holds significant importance for intensive care unit physicians and nurses daily, as it directly impacts the formulation of clinical decisions. The severity of trauma significantly influences the clinical prognosis of patients in traumatic coma (Sherer et al., 2008). While, prognostic assessment significantly influences daily clinical decisions for neuro-intensive care unit (NICU) physicians and nurses, particularly in managing nontraumatic coma patients. Currently, prognosis relies on key factors such as age, Glasgow Coma Scale (GCS) score, imaging findings, central nervous system signs, systemic indicators, and the underlying etiology (Bansal et al., 2005, Horsting et al., 2015, Rossetti et al., 2016). Knowledge of diverse comatose patient prognoses is essential for early intervention in family communication, treatment decisions (active treatment or end-of-life care), and monitoring patient indicators.

Electroencephalography (EEG) provides a dynamic representation of brain neuronal activity in both temporal and spatial domains, directly reflecting brain activity and metabolism that are not easily observable through imaging or biochemical markers (Rossetti et al., 2016, Grippo et al., 2017, Fahy and Chau, 2018, Sun et al., 2020). It is also a valuable tool to assess the depth of nontraumatic coma patients and the degree of brain function damage, which holds significant value in prognostic determination (Ardeshna, 2016, Goenka et al., 2018). The operational complexity of existing prognostic models, including sophisticated elements such as raw EEG patterns and reactivity, requires specialized neurophysiological knowledge. Recently, a novel neurofunctional monitoring tool known as quantitative EEG (QEEG) has emerged, representing a pioneering approach in the field of neurological assessment (Hwang et al., 2022). QEEG provides a more straightforward strategy which is a technique that compresses a large amount of raw EEG data from hours or even a whole day into a single frame for mapping and displaying the temporal and spatial distribution of the EEG (Haider, Esteller et al. 2016). It involves the analysis of EEG signals in both the frequency and time domains, culminating in the visualization of a patient's functional brain state through simplified intuitive trend plots that are easy to interpret (Hwang et al., 2022). QEEG is not only employed for the detection and diagnosis of epileptic seizures but also serves as a valuable tool for evaluating the degree of cerebral dysfunction in neurocritical patients. QEEG is utilized for the assessment of coma, cerebral perfusion, cerebral metabolism, and sedation depth and ultimately to determine the prognosis of the patient (Alkire, 1998, Haider et al., 2016, Tian et al., 2022). QEEG has the characteristics of high integration information, easy indexing, and convenient clinical interpretation, which is an important guide for the prognostic assessment of comatose patients (Haider et al., 2016). In contrast, continuous EEG requires continuous monitoring, requiring neurophysiologists to carefully go over each page of raw EEG to identify abnormalities. QEEG does not require physicians and nurses to delve into the technical details of EEG, although it does require clinicians to understand the basic concepts of EEG. It can help physicians and nurses understand the pattern of a patient's EEG activity, which is a significant time cost savings (Citerio et al., 2017; Wang, Huang et al. 2021; Tian et al., 2022). However, it represents a crucial yet unresolved clinical query about the selection of metrics that can yield optimal predictive efficacy.

In this study, we used the least absolute shrinkage and selection operator (LASSO) logistic regression analysis with QEEG and clinical indicators to develop a prognostic predictive model for nontraumatic coma patients three months after discharge. Developing straightforward prognostic instruments in NICU for clinicians would be beneficial, particularly in ascertaining the prognosis of various non-traumatic coma patients.