Volume 01 · Issue 03 · 2025

To the Editor:

Neurological disorders such as Alzheimer’s disease, Parkinson’s disease, and epilepsy pose significant challenges to public health owing to their complex pathophysiology. Early detection and accurate diagnosis are critical for effective intervention; however, traditional diagnostic methods often fall short in terms of sensitivity and specificity. Machine learning (ML) has shown great promise for overcoming these limitations by analyzing large-scale datasets, including neuroimaging and genomic data, to enhance diagnostic and predictive accuracy. This review explored the applications of ML in the prediction and diagnosis of Alzheimer’s disease, Parkinson’s disease, and epilepsy. We reviewed the principles of the ML algorithms, their performance in multimodal data analysis, and their potential for streamlining diagnostic workflows. Additionally, we discuss key challenges, such as model interpretability, data integration, and clinical adoption of ML technologies. Our goal was to highlight how ML can transform the prediction, diagnosis, and management of neurological disorders and ultimately improve patient outcomes.

Electrical stimulation has been used to influence the nervous system since ancient times, with early examples including the Egyptian use of electric fish for pain relief. The use of electrical stimulation to improve the understanding of brain regions and their functions advanced significantly in the 19th century, with experiments by Fritsch, Hitzig, and Bartholow, culminating in the development of techniques such as electrocorticography by Horsley. The field evolved with the development of stereotactic surgery in the 20th century, which allowed for the precise targeting of brain regions to treat movement disorders and chronic pain. Despite the advent of pharmacological treatments such as levodopa for Parkinson’s disease, surgical approaches such as thalamotomy and deep brain stimulation (DBS) remain crucial for patients who are unresponsive to medication. The refinement of DBS, particularly targeting the thalamus and subthalamic nucleus, has revolutionized the treatment of movement disorders and resulted in the widespread adoption of this technique. In contrast, early psychiatric treatments were typically rudimentary and driven by the dire conditions of mental asylums rather than scientific understanding. DBS has also been extended to the treatment of psychiatric disorders. Early studies conducted in the late 1990s and the early 2000s showed promising results for conditions such as obsessive-compulsive disorder. However, evidence of its efficacy in psychiatric conditions remains mixed, and ongoing research is required to establish its role. Ethical considerations are of paramount importance in this field, particularly concerning the safety, informed consent, and long-term effects of DBS. In contrast to movement disorders, psychiatric conditions experience challenges such as unclear biomarkers, subjective diagnoses, and unpredictable treatment responses, rendering DBS development in this area more complicated.

Extracellular vesicles (EVs) are defined as vesicles formed from lipid bilayers secreted by various cells. These EVs act as mediators between cells and transport proteins, genetic material, and lipids. They display significant heterogeneous and contain cell-specific markers. In the central nervous system, particularly in the brain, EVs are secreted primarily by neurons, astrocytes, microglia, and oligodendrocytes. Their involvement in neurological diseases has been extensively investigated. Various aspects of EVs have been examined to understand disease progression and modulation, particularly in neurodegenerative neurological diseases. In neurodegenerative diseases, EVs carry pathological proteins, facilitating protein spreading and neuroinflammation. In Alzheimer’s disease (AD), these vesicles carry Tau and amyloid-β proteins, causing their deposition and increasing the inflammatory response. Neuron-derived extracellular vesicles (NDEVs) have diagnostic potential as they contain pathological proteins central to the disease. Their observation in bodily fluids such as plasma and serum adds an advantage to their use as minimally invasive biomarkers. The levels of NDEV proteins can be measured to distinguish patients with AD from healthy controls. In addition, they can be used to predict disease development 10 years prior to clinical onset. Comparative analyses of the micro ribonucleic acids (miRNAs) and lipids present in NDEVs between patients with AD and controls have been performed. Many researchers have suggested that NDEVs could be utilized as AD diagnostic biomarkers because of their high potential for biomolecular differences.

Stroke is a leading cause of mortality and disability worldwide, and current therapeutic strategies are often limited in efficacy and scope. Gene therapy has emerged as a promising intervention that targets molecular pathways to mitigate neuronal damage and inflammation and enhance recovery. This study provides a comprehensive review of the recent advances, challenges, and future directions for the application of gene therapy in stroke management. A narrative review of the literature was conducted to evaluate the current status of gene therapy in stroke patients. Anti-inflammatory therapies using micro ribonucleic acids (miRNAs) and genetic targets have demonstrated efficacy in reducing inflammation and neuronal apoptosis in preclinical models. Delivery innovations, such as exosome-mediated systems and nanoparticle-based carriers, have enhanced the precision of gene therapy and the ability to cross the blood-brain barrier. Neuroprotective strategies, including vascular endothelial growth factor (VEGF) and heme oxygenase-1 (HO-1) gene delivery, showed promise in promoting recovery and reducing ischemic damage. Combined genetic and pharmacological approaches, such as clustered regularly interspaced short palindromic repeats (CRISPR)/dCas9-mediated Sirt1 upregulation, exhibit synergistic effects and improve therapeutic outcomes. However, most studies have focused on ischemic stroke with limited exploration of hemorrhagic stroke and its long-term safety profile. Key challenges include translational gaps between animal models and human physiology, complexity of stroke subtypes, off-target effects, and resource-intensive trial requirements. Ethical concerns such as informed consent, patient selection, and access disparities further complicate clinical applications.

Cathodal transcranial direct current stimulation (c-tDCS) is a safe and effective treatment for drug-resistant epilepsy (DRE). Examining changes in c-tDCS-induced functional connectivity (FC) is crucial for predicting treatment outcomes. Previous studies have estimated connectivity networks exclusively in awake patients, despite evidence that altered thalamocortical connectivity during sleep unmasks epileptic activity and may provide greater insight. The purpose of this study was to determine if FC analysis during sleep stages can provide a better assessment of tDCS effects.

Electroencephalogram-based FC network changes were assessed following c-tDCS in five adult patients with DRE treated at a tertiary epilepsy center, all of whom exhibited measurable clinical responses. Twenty-four-hour electroencephalogram (EEG) data were collected before and immediately after tDCS during the awake and non-rapid eye movement (NREM) sleep stages. FC changes were compared between the awake and sleep stages and correlated with the anti-epileptic effects of tDCS.

Significant FC alterations were observed in four of the five patients. Changes in FC during the awake state were noted in two patients, whereas a consistent pattern of increased FC during NREM sleep was evident in three of the five patients.

This study presents a novel evaluation of FC networks during NREM sleep to explore the anti-epileptic effects of tDCS. This pilot study provides an important foundation for understanding the impact of c-tDCS on FC networks during NREM sleep and its relation with anti-epileptic outcomes.

Frontal lobe epilepsy (FLE), the second most common refractory focal epilepsy, of which focal cortical dysplasia (FCD) is a common etiology. Electrophysiological analysis offers crucial insights into FCD subtype differentiation. Thus, we aim to evaluate scalp electroencephalography (EEG) and stereoelectroencephalography (SEEG) features, clinical characteristics, and postsurgical outcomes in a uniform series of patients with pathologically confirmed FCD type I or II who underwent FLE surgery.

Thirty-four consecutive patients were retrospectively analyzed. The interictal and ictal EEG and SEEG features, clinical characteristics, and surgical outcomes were evaluated.

Six interictal EEG patterns were identified. Focal continuous fast epileptiform discharges were only encountered in FCD II, with a high specificity of 100%. Continuous periodic spikes were predominantly observed in FCD II, with a high specificity of 94.4%; while repetitive discharges were largely observed in FCD I, with a specificity of 94.4%. Patients with a shorter epilepsy duration were more likely to show continuous irregular slowing (P = 0.034). Five EEG ictal onset patterns (IOPs) were identified. Spikes and slow waves were observed only in the FCD I group, whereas alpha activity was observed only in the FCD II group. Four interictal SEEG patterns were identified. Continuous periodic spikes were most frequently observed in the FCD II group (P = 0.044). Seven SEEG IOPs values were identified. Slow-wave or baseline shift followed by low-voltage fast activity (LVFA) and delta brush were observed only in FCD I. In contrast, preictal spiking followed by LVFA was mostly observed in the FCD II. Fast activity was observed only in the FCD I group. The average follow-up time was 1.8 years, and favorable surgical outcomes were observed more often in patients with FCD II (66.7%) than in those with FCD I (44.4%) but without significance (P = 0.315).

FCD subtypes share distinct EEG and SEEG signatures, and some special patterns may be indicative of specific subtypes.