Neurofeedback for autism spectrum disorder: Current evidence, challenges, and future directions

Abstract

Neurofeedback therapy (NFT) has emerged as a promising noninvasive intervention for autism spectrum disorder (ASD), targeting core symptoms such as social communication deficits and emotional dysregulation. This editorial synthesizes findings from recent studies, including Wang et al’s retrospective analysis (2025), which reported improvements in Social Responsiveness Scale and Aberrant Behavior Checklist scores following NFT combined with conventional therapy. Mechanistically, NFT may modulate prefrontal gamma-band activity, enhances neuroplasticity in social brain networks (e.g., default mode network, a brain network involved in social cognition), and optimizes cognitive processing via event-related potential changes (e.g., shortened P300 latency). Emerging trends include hybrid approaches (e.g., NFT with repetitive transcranial magnetic stimulation and artificial intelligence-driven protocols). However, challenges persist in protocol standardization, long-term efficacy validation, and biomarker identification. Future research must prioritize large-scale randomized trials, neuromarker discovery, and individualized protocols to establish NFT as a viable component of precision psychiatry for ASD.

Key Words: Autism spectrum disorder; Neurofeedback; Social skills; Emotion regulation; Electroencephalogram; Gamma oscillations; Prefrontal cortex; Neuromodulation; Precision medicine; Repetitive transcranial magnetic stimulation

Core Tip: Neurofeedback therapy (NFT), a non-invasive technique using real-time electroencephalography feedback to self-regulate brain activity, may offer benefits for neural dysregulation in autism spectrum disorder. However, current evidence is limited by small sample sizes and protocol variability. While NFT could potentially enhance neuroplasticity and complement behavioral therapies, its clinical application requires validation through rigorous randomized controlled trials. Future work should focus on personalizing protocols based on biomarkers and improving accessibility via telehealth solutions.

INTRODUCTION

Autism spectrum disorder (ASD) is a neurodevelopmental condition characterized by deficits in social communication and restricted behaviors[1]. Conventional interventions such as applied behavior analysis (ABA) often fail to address underlying neural dysregulation. Neurofeedback therapy (NFT), which enables self-regulation of brain activity via real-time electroencephalography (EEG) feedback, has emerged as a potential adjunctive approach. Wang et al[2] reported improvements in social responsiveness following NFT combined with conventional therapy in 2025. This editorial critically evaluates current evidence, mechanistic insights, and future directions for NFT in ASD, emphasizing methodological limitations[1,2].

CURRENT EVIDENCE AND LIMITATIONS

NFT shows potential as an intervention for ASD, with studies reporting improvements in core symptoms. For instance, NFT combined with conventional therapy has been associated with significant enhancements in social responsiveness, as measured by tools such as the Social Responsiveness Scale (SRS), and in cognitive functions, such as improved performance on executive function tasks such as the Flanker test[2,3]. These behavioral gains are often supported by electrophysiological changes, including reductions in P300 latency, which may indicate accelerated neural processing[4,5]. Mechanistically, NFT is thought to modulate prefrontal gamma-band activity and strengthen connectivity within social brain networks such as the default mode network (DMN), promoting neuroplasticity[6-8].

However, the evidence remains preliminary due to significant methodological limitations. Most studies have small sample sizes (often under 50 participants), which limits statistical power and generalizability[2,3]. Furthermore, protocols vary widely in terms of session duration, frequency, and target EEG frequencies, creating heterogeneity that hinders cross-study comparisons and meta-analytic synthesis[5]. A short-term focus is another constraint, as most trials lack long-term follow-up beyond six months, leaving the durability of effects uncertain[2]. The absence of double-blind designs in many studies also introduces risks of expectation bias as improvements may be influenced by non-specific factors such as placebo effects[5]. Future research must address these gaps through larger, standardized, and long-term trials to definitively establish NFT's efficacy for ASD[2,5] (Table 1).

Table 1 Key neurofeedback therapy findings and limitations[9-13].

Domain	Key findings	Effect size/Latency change	Limitations
Social communication	SRS scores enhance	15% reduction	Small samples; no active sham
Emotion regulation	ABC scores decrease; P300 latency decrease	∆41.95 ms (P < 0.05)	Short-term follow-up
Cognitive function	Executive function (Flanker test) enhance	∆1.59 (P < 0.001)	High-functioning ASD bias

SRS: Social Responsiveness Scale; ABC: Aberrant Behavior Checklist; ASD: Autism spectrum disorder.

MECHANISTIC INSIGHTS: FROM NEURAL CIRCUITS TO BEHAVIOR

NFT operates by enabling patients to voluntarily self-regulate brain activity through visual or auditory feedback derived from real-time EEG signals. The key mechanisms underpinning its efficacy in ASD include the following.

Prefrontal-cortex modulation and gamma-band training

NFT frequently targets the dorsolateral and medial prefrontal regions, which are critically involved in executive function and social cognition. Abnormal gamma-band activity (30-80 Hz) in these areas—often associated with excitatory/inhibitory imbalances in ASD—can be normalized through targeted training. Wang et al’s protocol[2], which incorporated cartoon-based attention tasks to enhance gamma activity, aligns with previous studies demonstrating improved social responsiveness post intervention[8]. For instance, Wang et al[2] reported that beta-band NFT increased functional connectivity within the default mode and salience networks, reducing static network variability and concomitantly improving social behaviors[4] (Table 2).

Table 2 Emerging innovations in neurofeedback therapy and their clinical applications.

Innovation domain	Key technological advance	Clinical application	Evidence strength
Hybrid neuromodulation	rTMS priming + NFT reinforcement	Enhanced reduction of repetitive behaviors and hyperactivity in ASD	Controlled trials[6]
Artificial intelligence integration	Machine learning-driven real-time parameter adjustment	Personalized training intensity based on moment-to-moment brain states	Platform development[3]
Telehealth expansion	Portable EEG devices with cloud connectivity	Home-based treatment for ASD with remote supervision	Market report[2]
Biomarker personalization	EEG neuromarkers (e.g., P300, gamma ratios) for prediction	Protocol selection based on individual neurophysiological profiles	Research validation[4]

rTMS: Repetitive transcranial magnetic stimulation; NFT: Neurofeedback therapy; ASD: Autism spectrum disorder; EEG: Electroencephalography.

Neuroplasticity and functional connectivity

NFT promotes experience-dependent synaptic plasticity within social brain networks (e.g., DMN, mirror neuron system) by reinforcing the desired EEG oscillatory patterns. Combining NFT with repetitive transcranial magnetic stimulation (rTMS) enhanced EEG gamma activity and reduced repetitive behaviors, suggesting the strengthening of neural pathways subserving behavioral flexibility[6]. Additionally, NFT mitigates ASD-related connectopathy—characterized by local overconnectivity and long-range underconnectivity—by fostering increased integration between the frontal and posterior regions[4].

Event-related potential correlates of cognitive improvement

NFT induces measurable changes in event-related potentials (ERPs), reflecting optimized neural resource allocation and processing efficiency. Slow cortical potential training shortened P300 latency (331.41 ms post-training vs 373.36 ms baseline), indicating accelerated emotional stimulus evaluation[5]. Moreover, changes in late positive potential amplitude correlated with improved positive emotion processing (r = 0.53), underscoring enhanced emotional regulation capacity[5].

CLINICAL EFFICACY AND CHALLENGES

NFT has demonstrated potential efficacy in addressing core symptoms of ASD, such as social communication deficits and emotional dysregulation. Studies including a retrospective analysis by Wang et al[2] indicate that NFT, particularly when combined with conventional rehabilitation, can lead to significant improvements in standardized metrics such as the SRS and Aberrant Behavior Checklist scores, with some reports suggesting superior outcomes compared to conventional therapy alone[2]. Meta-analytical evidence supports these findings, showing that NFT may have positive effects on social communication and behavior, although effect sizes vary[3].

Therapeutically, NFT is thought to modulate aberrant neural circuits by normalizing electrophysiological imbalances, such as excessive theta/beta ratios, and enhancing gamma-band activity in prefrontal regions. These changes are associated with improved functional connectivity within critical networks such as the DMN, which is involved in social cognition. For example, reductions in P300 latency following NFT correlate with accelerated emotional stimulus evaluation[2,5]. These mechanisms may promote neuroplasticity, leading to sustained benefits. Long-term observations include maintained improvements in executive functions (e.g., performance on Flanker tests) and working memory for several months post-intervention[2]. Additionally, caregivers often report enhanced social initiation and reduced emotional outbursts[9]. NFT is generally well-tolerated in pediatric populations, with adverse effects typically being mild and transient, such as temporary headaches[4].

However, the evidence base faces significant challenges. A major limitation is the lack of protocol standardization, with substantial variation in session duration, training frequency, and target neural parameters across studies, which hinders comparison and replication[3]. Most research involves small sample sizes (often under 50 participants) and lacks long-term follow-up data beyond six months, leaving questions about the durability of effects and potential need for booster sessions[2,5]. The high cost of clinical NFT systems and need for trained practitioners limit accessibility, while emerging home-based devices raise concerns about data security and treatment fidelity without professional supervision[8]. Furthermore, mechanistic uncertainties persist; it remains unclear whether behavioral improvements are directly due to neural modulation or influenced by non-specific factors such as placebo effects[2,3]. Finally, the generalizability of findings is constrained as most studies focus on high-functioning ASD individuals, with efficacy in low-functioning or non-verbal populations remaining largely unexplored[2,5]. Future research requires large-scale, sham-controlled trials and biomarker validation to establish NFT's role in ASD management.

COMPARATIVE ADVANTAGES OVER TRADITIONAL THERAPIES

NFT demonstrates several distinct advantages over conventional treatment approaches, primarily through its capacity to guide self-regulation of neural function via real-time EEG feedback. In contrast to pharmacotherapy (e.g., antipsychotics or stimulants), NFT is non-invasive and medication-free, thereby avoiding drug-related adverse effects such as weight gain, sedation, or cardiovascular risks. Treatment-emergent adverse events (e.g., transient headache or fatigue) are generally mild and infrequent, underscoring NFT’s suitability for pediatric populations[2,5].

A key strength of NFT lies in its personalized intervention approach. Protocols are tailored to individual EEG patterns—such as elevated theta/beta ratio or excessive gamma activity—enabling precise neuromodulation of dysfunctional neural circuits. This individualized approach contrasts with traditional behavioral therapies (e.g., ABA), which often employ generalized methodologies; additionally, NFT dynamically adapts parameters based on real-time neural activity[3].

Furthermore, NFT exhibits significant complementary effects when integrated with existing interventions. When combined with behavioral therapy, it may enhance neuroplasticity and facilitate skill acquisition by priming neural networks. In comparison, pharmacotherapy often targets symptomatic relief without promoting sustained neural adaptation[2,5].

In terms of treatment durability, NFT-induced neuroplastic changes often yield sustained therapeutic benefits, potentially reducing the need for long-term intervention. This contrasts with behavioral techniques, which typically require continuous implementation, and pharmacotherapies, whose effects commonly diminish following discontinuation[5].

EMERGING TRENDS AND INNOVATIONS

The future advancement of NFT necessitates a multifaceted research agenda to establish efficacy, optimize protocols, and ensure clinical translatability. A primary imperative is the execution of large-scale, multi-site randomized controlled trials (RCTs) employing active sham-control conditions (e.g., placebo EEG feedback) to isolate NFT-specific effects from non-specific factors such as participant expectation. Such studies must enroll demographically and clinically diverse cohorts to enhance generalizability[2].

Concurrently, biomarker-driven personalization is essential, focusing on validating electrophysiological neuromarkers (e.g., P300 latency, gamma power spectral ratios) that predict individual treatment response. Integrating machine learning for real-time EEG analysis can enable closed-loop systems that adapt to neurophysiological states, maximizing outcomes[2,5]. Furthermore, hybrid approaches combining NFT with other neuromodulation techniques such as rTMS or transcranial direct current stimulation may yield synergistic effects by “priming” neural circuits, as evidenced by enhanced reductions in repetitive behaviors[3]. Improving accessibility via telehealth and affordable, portable EEG devices is crucial, although challenges regarding data security and privacy require solutions such as encryption protocols[3]. Longitudinal studies tracking outcomes over 1-5 years are required to identify optimal intervention windows and integrate NFT within multimodal frameworks (e.g., behavioral therapy, virtual reality), promoting real-world skill generalization[4].

These innovations collectively represent a potential approach toward more precise, accessible, and effective neuromodulation therapies that are increasingly integrated with digital health infrastructures and personalized through computational analytics[5].

FUTURE DIRECTIONS

The future advancement of NFT for ASD requires a multifaceted research agenda to establish efficacy and ensure clinical translatability. A primary imperative involves conducting large-scale, multi-site RCTs employing active sham-control conditions to isolate NFT-specific effects from non-specific factors while enrolling demographically and clinically diverse cohorts to enhance generalizability[2]. Concurrently, biomarker-driven personalization is essential, focusing on validating electrophysiological neuromarkers such as P300 latency or gamma power spectral ratios to predict individual treatment response. Integrating machine learning for real-time EEG analysis can enable closed-loop systems that adapt to neurophysiological states, optimizing outcomes[2,5]. Furthermore, hybrid approaches combining NFT with other neuromodulation techniques such as rTMS may yield synergistic effects by priming neural circuits, as evidenced by enhanced reductions in repetitive behaviors[3]. Improving accessibility via telehealth and affordable, portable EEG devices is crucial, although challenges regarding data security and privacy require solutions like encryption protocols[4]. Longitudinal studies tracking outcomes over 1-5 years are needed to identify optimal intervention windows and integrate NFT within multimodal rehabilitation frameworks, promoting real-world skill generalization[5] (Table 3).

Table 3 Key future research directions in neurofeedback therapy.

Research direction	Primary objectives	Key considerations & challenges
Large-scale RCTs	Establish causal efficacy, isolate NFT-specific effects from placebo	Requires significant funding, multi-site collaboration, development of credible sham protocols
Biomarker personalization	Identify predictive neuromarkers (e.g., P300, gamma), develop adaptive algorithms	High inter-individual variability, need for standardized measurement and analysis
Hybrid neuromodulation	Explore synergistic effects of NFT with rTMS/tDCS, prime neural circuits	Protocol optimization, safety of combined modalities, mechanistic understanding
Telehealth & home-based NFT	Increase accessibility, reduce costs, enable remote training	Device affordability and reliability, data security, user adherence, signal quality
Longitudinal & developmental	Assess durability of effects, identify critical intervention windows	Long-term funding, participant retention, controlling for confounding life factors
Multimodal integration	Enhance outcomes by combining NFT with behavioral therapy, OT, exercise, VR	Treatment coordination, cost, measuring specific contribution of each component

RCT: Randomized controlled trial; NFT: Neurofeedback therapy; rTMS: Repetitive transcranial magnetic stimulation; tDCS: Transcranial direct current stimulation; OT: Occupational therapy; VR: Virtual reality.

In summary, the future trajectory of NFT research hinges on rigorous efficacy trials, personalized protocol optimization through computational analytics, strategic combination with complementary neuromodulation techniques, and thoughtful implementation within accessible, integrated, and ecologically valid intervention frameworks. Addressing these priorities will be pivotal in translating NFT from a promising investigational tool into an established, evidence-based clinical intervention.

CONCLUSION

NFT represents a potential approach in ASD intervention, directly targeting the neural circuitry underpinning social-emotional and cognitive deficits. Wang et al’s study[2], alongside recent clinical trials and meta-analyses, underscores its potential to ameliorate core symptoms through mechanisms such as prefrontal gamma modulation, ERP optimization, and enhanced functional connectivity. Emerging innovations—including hybrid neuromodulation, artificial intelligence-driven protocols, and biomarker personalization—hold promise for further enhancing NFT’s efficacy and accessibility. However, addressing challenges related to protocol standardization, costs, and mechanistic clarity through collaborative, rigorous research is essential. As the field advances, NFT may evolve into a viable component of precision psychiatry for ASD, offering renewed hope for millions affected by this complex and heterogeneous disorder.