Alertness, a pivotal cognitive state affecting reaction time, decision-making, and overall mental acuity, has traditionally been assessed via subjective scales or time-intensive behavioral tests. These conventional assessments are often plagued by inherent drawbacks such as subjectivity, susceptibility to external interference, and a reliance on complex, laborious data collection protocols. The demand for a more streamlined, precise, and non-invasive evaluation tool has thus propelled researchers toward novel neuroengineering solutions that interface seamlessly with brain activity markers.

The latest study approached this challenge using optical stimulation to induce brainwave entrainment (BWE) across a wide range of frequencies, from the theta to gamma bands (4 to 48 Hz). Brainwave entrainment refers to the phenomenon where external rhythmic stimuli synchronize neural oscillations within the brain, effectively modulating brain states. The researchers exposed 40 human subjects to 12 discrete optical stimulation frequencies, each applied for a duration of 30 seconds, while recording real-time electroencephalogram (EEG) signals from the prefrontal cortex, specifically at the Fpz, Fp1, and Fp2 electrode sites.

The prefrontal cortex selection is particularly significant, given its established role in executive functions and maintaining vigilance and alertness. By focusing on these areas, the study sought to capture the nuanced electrophysiological signatures that correlate with fluctuations in alertness. Complementing the EEG data, subjective alertness evaluations were conducted through widely accepted measures, including the Karolinska Sleepiness Scale and the psychomotor vigilance test, providing a comprehensive behavioral correlate to the neural findings.

Central to the analysis was the extraction of nine distinct EEG features during the periods of BWE. These features, representing various brainwave power bands and entrainment characteristics, were scrutinized for their correlation with alertness states. For example, the β (beta) and γ (gamma) bands—commonly associated with focused attention and higher cognitive processing—were investigated for their differential expression relative to baseline states. Understanding which features are most predictive of alertness could pave the way for targeted neuromodulation therapies or assessment protocols.

In an impressive step beyond traditional statistical methods, the researchers integrated machine learning classifiers, including support vector machines, Naive Bayes, and logistic regression models, to create an automated system capable of distinguishing between different alertness levels based on real-time EEG input during BWE. This integration underscores the growing synergy between neuroengineering and artificial intelligence, facilitating more robust and scalable brain state detection frameworks.

Notably, among the various frequencies tested, the 30 Hz optical stimulation emerged as the most effective for inducing discernible brainwave entrainment changes linked to alertness variation. Particularly, three features—BWE intensity, β band power, and γ band power—demonstrated statistically significant differences between alertness states. Utilizing these parameters, the Naive Bayes model achieved outstanding classification performance, with an area under the curve (AUC) reaching 0.90, alongside high accuracy, sensitivity, and specificity values approximating 90%. Such results underscore the potential clinical and applied utility of this approach in real-world settings.

Perhaps unexpectedly, the researchers observed that the subjects’ alertness levels, as measured behaviorally and subjectively, did not significantly shift before and after the 30-second optical stimulation periods. This finding implies that while short-term optical stimulation effectively modulates EEG features predictive of alertness, it might not transiently alter the underlying alertness state itself. Hence, the protocol primarily serves as a diagnostic or assessment tool rather than an intervention intended to enhance alertness.

The broader implications of this work are manifold. The capacity to rapidly and non-invasively assess alertness via optical BWE combined with machine learning could revolutionize fields where human vigilance is paramount—from transportation safety to workforce performance monitoring. Moreover, this modality’s ease of implementation and quantifiability positions it as a prime candidate for integration into portable neurodiagnostic devices, wearable technology, or even consumer-grade applications tailored for cognitive state monitoring.

Furthermore, the study’s methodological rigor in correlating subjective scales, behavioral tests, and electrophysiological markers lays a robust foundation for future research. It invites exploration into longer stimulation durations, diverse stimulation frequencies, or multimodal sensory entrainment approaches that might also influence alertness levels transiently or chronically. The potential to personalize alertness assessment based on individual neural responsiveness to optical stimuli also represents an exciting frontier.

From a neuroscience perspective, the findings deepen our understanding of how rhythmic external stimuli interact with cortical oscillations related to vigilance and attention. The demonstrated significance of β and γ bands in reflecting alertness aligns with extant literature that associates these frequencies with active cognitive engagement. This convergence bolsters confidence in the neurobiotic validity of the proposed assessment technique.

While the study focuses on healthy adult participants, future research may expand to include clinical populations where alertness monitoring is critical, such as those with sleep disorders, neurological diseases, or cognitive impairments. Additionally, integrating this technique with other physiological measures like heart rate variability or pupillometry could enhance multidimensional assessment models, providing richer data for personalized alertness profiles.

In summary, this innovative approach of optical stimulation-induced brainwave entrainment assessed via sophisticated machine learning classifiers marks a significant stride in cognitive state monitoring technology. It bridges neurophysiological insights with cutting-edge artificial intelligence tools to deliver a rapid, objective, and practical alertness evaluation method. As the fields of neurotechnology and data science continue to intertwine, such interdisciplinary research foreshadows a future where mental states can be precisely measured and potentially modulated in everyday life.

The evolution of alertness assessment from subjective questionnaires and reaction-based tests to real-time EEG biomarkers interpreted through machine learning heralds a paradigm shift. This breakthrough not only enhances fundamental scientific understanding but also promises wide-reaching applications in safety, productivity, and health monitoring, ultimately contributing to improved quality of life and prevention of cognitive lapses.

As technology advances, coupling optical stimulation-induced brainwave entrainment with wearable EEG devices may render continuous alertness tracking feasible in practical environments. Such advancements could enable adaptive systems that respond dynamically to users’ cognitive states—alerting when vigilance drops or optimizing task demands for peak performance—revolutionizing human-machine interaction.

The study by Zhou, Tan, Wang, et al., represents a compelling example of how multidisciplinary collaboration can translate complex brain activity patterns into actionable data. Their work provides the scientific community and industry stakeholders with an invaluable roadmap for next-generation cognitive state monitoring protocols informed by neurophysiology and powered by artificial intelligence.

Subject of Research: Rapid and objective assessment of human alertness through optical stimulation-induced brainwave entrainment combined with machine learning analysis of EEG features.

Article Title: Alertness assessment by optical stimulation-induced brainwave entrainment through machine learning classification.

Article References:
Zhou, Y., Tan, Y., Wang, S. et al. Alertness assessment by optical stimulation-induced brainwave entrainment through machine learning classification. BioMed Eng OnLine 24, 100 (2025). https://doi.org/10.1186/s12938-025-01422-4

Image Credits: AI Generated

DOI: https://doi.org/10.1186/s12938-025-01422-4

Dr. Kaplan’s career trajectory is as unconventional as it is inspiring. From a foundation in journalism to her rigorous training as a clinical psychologist, her work is a testament to the power of interdisciplinary thinking. The genesis of her research focus lies in personal observations—specifically, witnessing her mother’s journey coping with chronic pain caused by a connective tissue disorder unveiled significant limitations in clinical assessment tools. Traditional research often occurs in controlled laboratory settings, failing to capture the intricate ebb and flow of symptoms and emotions experienced in real-world contexts, a discrepancy Dr. Kaplan set out to address.

Her engagement with silent meditation retreats further galvanized her inquiry, prompting a critical question: do the transformative experiences gained in controlled spiritual environments manifest authentically in the fabric of daily living? This curiosity propelled her to pursue doctoral studies under the mentorship of Dr. Matthias Mehl at the University of Arizona, a luminary in ambulatory assessment—an approach that embraces participants’ lived realities through real-time data collection.

At Emory University, Dr. Kaplan identified a crucial methodological bottleneck. Conventional instruments for gathering experiential data leaned heavily on rigid surveys or passive audio captures, missing the spontaneity and depth of personal narratives. To surmount these limitations, she collaborated with Santiago Arconada Alvarez, Co-Director of Mobile Apps at Emory’s AppHatchery, to develop Fabla—a secure, smartphone-based application designed to collect unstructured voice narratives from participants throughout their daily routines.

Fabla’s significance resides in its capability to harvest what Dr. Kaplan describes as “voice memos for research.” Unlike scripted surveys or clinical interviews, these voice memos capture candid, unsupervised reflections that reveal how subjects internalize and process therapeutic interventions in the moment. This innovative method has transformed the scope of qualitative data, creating a rich tapestry of subjective experience that conventional technologies have failed to access.

Though initially created with psychedelic-assisted therapy research in mind, Fabla’s adaptability has led to its adoption across a variety of health disciplines. More than a dozen research initiatives now leverage the platform to explore veteran rehabilitation during substance use treatment, patient narratives in epilepsy management, and even the psychological toll of burnout among healthcare providers. This broad uptake underscores a widespread demand within clinical research for tools that preserve the complexity and authenticity of human experience.

Dr. Kaplan’s work is uniquely timely, coinciding with rapid advances in artificial intelligence, particularly generative large language models capable of analyzing speech data on an unprecedented scale. These technologies offer enormous promise for accelerating research but simultaneously introduce a host of ethical challenges. How can researchers harness AI’s computational power without diluting the relational and contextual subtleties that imbue voice data with meaning?

The ethical dimension of voice analysis forms a critical pillar of Dr. Kaplan’s current research agenda. Automated systems can now detect emotional states, cognitive patterns, and even inferred psychological conditions directly from voice recordings. However, this transformative potential is tempered by concerns regarding privacy, consent, and the risk of eroding the nuanced intersubjective qualities of human communication—those shared meanings created through relational engagement, which are particularly vital in understanding psychedelic experiences.

Psychedelic therapy exemplifies these challenges. Patients often describe ineffable shifts in perception, identity, and consciousness—phenomena difficult to codify or quantify. Dr. Kaplan questions whether algorithmic interpretation might oversimplify or misrepresent these profound transformations. Her lab emphasizes maintaining reflexivity in research methods, acknowledging the limitations of technology and advocating for an ethical framework that respects the depth of subjective experience.

This reflective stance extends into Dr. Kaplan’s scientific philosophy. She recalls guidance from her clinical training director: “When in doubt, be a human.” This ethos sharply contrasts the increasing pressure in research environments to prioritize speed and productivity at the potential cost of care and depth. Her laboratory culture fosters mindfulness and prioritizes the human element, encouraging students and researchers to resist reductive trends and honor the complexity of lived experience.

Balancing technology with contemplative practice also informs Dr. Kaplan’s personal and professional life. She maintains an adapted Shabbat ritual with her partner, a weekly 24-hour period without technology that mirrors her investigation into how moments of stillness and presence can solidify insights from therapeutic interventions into sustainable life changes. Her research seeks to unravel what enables lasting integration of transformative experiences beyond isolated sessions.

Looking forward, Dr. Kaplan envisions a future where behavioral science and artificial intelligence converge, reshaping both academic research and applied mental health care. With the proliferation of voice-enabled AI assistants and mental health applications, there is an urgent need to develop standards for ethically capturing and analyzing voice data that preserve dignity and authenticity. Her work asks poignant questions: Can voice biomarkers complement traditional clinical assessments? What safeguards must be embedded to ensure that the efficiency of automation does not eclipse the therapeutic power embedded in being genuinely heard?

The Genomic Press interview with Dr. Deanna M. Kaplan is part of the “Innovators & Ideas” series, spotlighting transformative figures at the forefront of scientific progress. By weaving professional achievements with deeply human stories, the series offers readers not only insight into cutting-edge discoveries but also into the motivations and philosophies driving today’s research revolutionaries. Dr. Kaplan’s story and her visionary work epitomize the potential for technology to both illuminate and preserve the intricacies of human experience in clinical science.

Subject of Research: People

Article Title: Deanna M. Kaplan: Listening to daily life: exploring speech data, shared meaning, and generative artificial intelligence (AI) in clinical science

News Publication Date: 3 June 2025

Web References:
DOI: 10.61373/pp025k.0023

Image Credits: Dr. Deanna M. Kaplan

Keywords: voice technology, clinical psychology, psychedelics, qualitative data, Fabla app, ambulatory assessment, artificial intelligence, ethical voice analysis, generative AI, human experience, behavioral science, mental health technology