Neurofeedback is a type of biofeedback training where individuals learn to control and normalize patterns of brain activity using real-time displays of neural signals.
By teaching self-regulation of brain function, neurofeedback can be a non-pharmacological treatment approach for various mental health conditions like ADHD, anxiety, depression, and more.
Ging-Jehli, N. R., Painter, Q. A., Kraemer, H. A., Roley-Roberts, M. E., Panchyshyn, C., deBeus, R., & Arnold, L. E. (2023). A diffusion decision model analysis of the cognitive effects of neurofeedback for ADHD. Neuropsychology. Advance online publication. https://doi.org/10.1037/neu0000932
Key Points
- Neurofeedback (NF) treatment for ADHD seems to improve the efficiency of processes associated with integrating auditory stimulus information more than control treatment.
- NF also improved context sensitivity, making responses more consistent irrespective of trial type, particularly for auditory trials.
- Before treatment, children with ADHD showed deficits mainly in integrating stimulus information compared to healthy children.
- The diffusion decision model analysis suggests NF improves the specific cognitive components impaired in ADHD.
Rationale
Attention-deficit hyperactivity disorder (ADHD) is a common childhood neurodevelopmental disorder characterized by inattention, hyperactivity, and impulsivity (American Psychiatric Association, 2013).
Medications have limited efficacy, with up to a third of children not responding, highlighting the need for alternative treatments (Adler et al., 2006).
Neurofeedback (NF) is a promising non-pharmacological intervention where individuals learn to modulate brain activity patterns. However, NF’s effects on cognition are unclear. Understanding the cognitive mechanisms of NF is critical for treatment planning and personalization.
This study comprises a secondary analysis of a large double-blind randomized controlled trial (RCT) on NF for ADHD (Neurofeedback Collaborative Group, 2021).
Although the primary outcome of ADHD symptom improvement showed nonspecific gains for all children, moderator analyses indicated NF efficacy depended on comorbidities. Additionally, baseline cognitive signatures predicted NF response (Ging-Jehli et al., 2023).
Given the mixed results and NF’s conceptualization as a reinforcement learning treatment (Lubianiker et al., 2022), investigating NF’s cognitive effects is vital.
This study aimed to assess whether NF improves the efficiency of information processing, a core deficit in ADHD (Ging-Jehli et al., 2021), using computational modeling. The diffusion decision model (DDM) was applied to data from a continuous performance test to quantify latent cognitive components before and after treatment.
The DDM is a commonly applied model that describes how people make decisions by providing a simplified blueprint of the decision-making process. It quantifies performance into components with established psychological meaning that can be studied separately, including response cautiousness, the efficiency of information integration, context sensitivity, response bias, and non-decision time.
Method
Materials/Instruments:
- Integrated visual and auditory continuous performance test (IVA2-CPT): A computerized test that presents visual or auditory stimuli to assess sustained attention/impulse control where participants respond to “go” trials and withhold responses to “no-go” trials.
- Children’s Interview for Psychiatric Syndromes (ChIPS): A structured diagnostic interview administered to children to determine eligibility based on ADHD diagnosis and comorbid mental health disorders according to the DSM-5 criteria.
- Conners Parent Rating Scales: Standard rating scales completed by parents/teachers to measure symptoms of hyperactivity, impulsivity, and/or inattention used as outcome measures in the RCT.
Procedure
Secondary analysis of a double-blind RCT; diffusion decision model (DDM) analysis
The ADHD group received 38 sessions of NF or controlled treatment with assessments at baseline, mid-, and end-treatment. Controls had one assessment.
Sample
- 133 children aged 7-10 with ADHD (78 boys, 54 girls)
- 57 healthy controls (31 boys, 26 girls); controls were one year older than ADHD groups
- No significant demographic differences between ADHD treatment groups
Statistical Measures
- Linear mixed models on DDM parameters (efficiency of information integration phi v; context sensitivity cv); follow-up analyses calculating within-group changes
- Kenward-Roger approximation for multiple comparisons
- Follow-up analyses used Hedges’ g effect sizes
Results
The key analysis showed the efficiency of integrating auditory information (phi v) improved significantly more over time with NF compared to control treatment, supporting the hypothesis.
Additionally, context sensitivity (cv) improved more over time with NF, indicating more consistent responses across trial types, especially for auditory trials.
Comparing ADHD to healthy controls at baseline revealed significantly lower phi v in ADHD, confirming integration inefficiency as a deficit before treatment. NF also reduced latency of non-decisional processes like encoding and motor execution (Ter) whereas control treatment showed no Ter changes.
However, response cautiousness (a) and bias (z/a) did not significantly differ between groups over time.
Insight
Applying computational modeling provided valuable insights into NF’s effects on cognition.
The findings suggest NF may be beneficial by targeting the specific deficits present in ADHD, namely improving efficiency and consistency of information processing.
The significant gains in auditory performance align with previous evidence showing auditory deficits in ADHD (Ging-Jehli et al., 2022).
As auditory stimuli unfold sequentially and require sustained attention, unlike visual input allowing a “second look”, improving auditory processing abilities could promote attention and reduce distraction.
The reductions in non-decision time could also indicate faster encoding and response readiness resulting from NF training.
Together, these cognitive changes likely contribute to behavioral improvements in attention and impulse control that characterize ADHD.
By focusing analyses on component processes, this study illustrates how NF may remediate the underlying pathophysiology as opposed to just alleviating symptoms. Additionally, some ADHD medications target similar cognitive mechanisms (e.g. drift rates), further validating deficient information integration as a core dysfunction (Ging-Jehli et al., 2021).
Applying computational methods could therefore clarify the neurobiological effects of various ADHD treatments.
Strengths
This research had several key strengths:
- Use of computational modeling (DDM analysis) to quantify latent cognitive processes and determine which components are affected by NF training. This advanced method provides more specific insights compared to analyzing performance summary statistics.
- Leveraging data from a large-scale, double-blind, placebo-controlled RCT, which rigorously tests the unique effects of NF while controlling other variables. The RCT design demonstrates causality more robustly than correlation studies.
- Focusing analyses on an objective, performance-based cognitive task not used in NF training. Finding transfer effects to untrained tasks strengthens conclusions about NF’s cognitive benefits.
- Comparing children with ADHD to a healthy control group at baseline. Documenting poorer efficiency of information integration verified this is an area of dysfunction in ADHD that NF could potentially remediate.
- Examining both visual and auditory modalities of the continuous performance test. Being able to detect auditory-specific improvements implies NF may restore sensory processing deficiencies underlying ADHD.
Limitations
However, this study also had some limitations:
- The analysis relied solely on one cognitive test, the IVA2-CPT. Incorporating multiple neuropsychological tasks assessing various functions would give a more comprehensive picture of NF’s effects.
- Participants were restricted to children aged 7-10 years with specific ADHD diagnoses and IQ criteria. The results may not generalize to other ages, ADHD presentations, or individuals with intellectual disabilities.
- There was no long-term follow-up testing to determine if cognitive changes persisted beyond treatment completion. It is unclear if gains would be maintained or would decay without ongoing NF.
- The ADHD group contained heterogeneity with regards to comorbidities. Although subgroupmoderator analyses were done in the RCT (Roley-Roberts et al., 2023), this secondary analysis did not differentiate comorbidity profiles, which could obscure some effects.
- The lack of an active control condition limits causal claims about specific ingredients underlying cognitive improvements. Nonspecific factors like interacting with technology, receiving attention from researchers, or expectations could also affect performance.
Implications
Even with the limitations, these results have useful real-world implications.
Most importantly, they indicate NF may work by fixing the underlying cognitive problems in ADHD instead of just improving symptoms. This means doctors can use neurocognitive test results to predict who might benefit most from NF and target certain thinking processes during training.
Patients with more difficulty efficiently integrating information and paying auditory attention seem especially likely to improve with NF. The thinking patterns spotted through computer modeling could be used along with symptom ratings to personalize ADHD treatment.
The findings also show auditory processing should get more focus. NF protocols may need to emphasize sound-based tasks more to optimize gains.
Sounds surround us at school, home, and socially, so better auditory skills could really help functioning and quality of life. Doctors should start checking for auditory issues, too, since these are often overlooked.
Additionally, the results support NF as an alternative or extra therapy to stimulant meds. Showing NF impacts brain function like medication does is promising for NF as a non-drug option targeting underlying ADHD issues. Parents and physicians have wanted more evidence-based choices, given many patients struggle with medication side effects and taking it as prescribed.
Lastly, using computer modeling here is an innovative way to study how treatments work. Applying similar testing to understand psychotherapy or new drugs could identify their effects on thinking.
Overall, adding computational methods to clinical trials may speed up the development of personalized therapies.
References
Primary reference
Ging-Jehli, N. R., Painter, Q. A., Kraemer, H. A., Roley-Roberts, M. E., Panchyshyn, C., deBeus, R., & Arnold, L. E. (2023). A diffusion decision model analysis of the cognitive effects of neurofeedback for ADHD. Neuropsychology. Advance online publication. https://doi.org/10.1037/neu0000932
Other references
Adler, L. A., Reingold, L. S., Morrill, M. S., & Wilens, T. E. (2006). Combination pharmacotherapy for adult ADHD. Current Psychiatry Reports, 8(5), 409–415. https://doi.org/10.1007/s11920-006-0044-9
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). https://doi.org/10.1176/appi.books.9780890425596
Ging-Jehli, N. R., Kraemer, H. C., Eugene Arnold, L., Roley-Roberts, M. E., & deBeus, R. (2023). Cognitive markers for efficacy of neurofeedback for attention-deficit hyperactivity disorder–personalized medicine using computational psychiatry in a randomized clinical trial. Journal of Clinical and Experimental Neuropsychology, 45(2), 118-131. https://doi.org/10.1080/13803395.2023.2206637
Ging-Jehli, N. R., Ratcliff, R., & Arnold, L. E. (2021). Improving neurocognitive testing using computational psychiatry-A systematic review for ADHD. Psychological Bulletin, 147(2), 169-231. https://doi.org/10.1037/bul0000319
Lubianiker, N., Paret, C., Dayan, P., & Hendler, T. (2022). Neurofeedback through the lens of reinforcement learning. Trends in Neurosciences, 45(8), 579-593. https://doi.org/10.1016/j.tins.2022.03.008
Roley-Roberts, M. E., Pan, X., Bergman, R., Tan, Y., Hendrix, K., deBeus, R., Kerson, C., Arns, M., Ging Jehli, N. R., Connor, S., Schrader, C., & Arnold, L. E. (2023). For which children with ADHD is TBR neurofeedback effective? Comorbidity as a moderator. Applied Psychophysiology and Biofeedback, 48(2), 179-188. https://doi.org/10.1007/s10484-022-09575-x
The Neurofeedback Collaborative Group. (2021). Double-blind placebo-controlled randomized clinical trial of neurofeedback for attention-deficit/hyperactivity disorder with 13 month follow-up. Journal of the American Academy of Child & Adolescent Psychiatry, 60(7), 841-855. https://doi.org/10.1016/j.jaac.2020.07.906
Keep Learning
- How might the cognitive improvements from NF training translate to real-world behavioral and functional changes? What evidence would be needed to demonstrate ecological validity?
- Should computational modeling be incorporated into all clinical trials testing psychiatric treatments to clarify the neurological mechanisms underlying response? What are the barriers to implementing this on a wider scale?
- Could the information integration deficit be a cross-diagnostic marker transcending categorical diagnoses like ADHD? Might this domain dysfunction also operate in mood or anxiety disorders?
- What patient factors could moderate or mediate NF’s effects on cognition and symptoms? For example, how might comorbidities influence treatment response?
- If deficits in information sampling efficiency reflect immature brain development in ADHD, do the changes from NF represent accelerated maturation? Or does NF work through fundamentally different processes?
