Multi-Arm Phase 2 Trial Fails to Show Efficacy for Cognitive Programs in Treating Long COVID

In a recently published phase 2 trial, findings showed that use of 3 remotely delivered interventions–an online cognitive training, a structured cognitive rehabilitation program, and transcranial direct current stimulation (tDCS)–failed to show effectiveness as a treatment option for long COVID. While all 3 arms showed some improvements in self-reported symptoms, the study highlights the difficulties with treating long COVID, a condition for which there are no successful pharmacological interventions.¹

Otherwise known as the RECOVER-NEURO trial (NCT05965739), the study tested 3 remotely administered interventions among a larger group of 5 arms. The study arms included an (1) active comparator (video games and puzzles), the control condition for all interventions; (2) BrainHQ; (3) PASC-CoRE + BrainHQ; (4) tDCS-active + BrainHQ; and (5) tDCS-sham + BrainHQ, the specific control arm for the tDCS-active arm.

For context, BrainHQ is a computerized, adaptive cognitive-training program focused on improving memory, attention, and processing speed through daily, gamified exercises, whereas PASC-CoRE is a structured, therapist-led cognitive rehabilitation program—designed specifically for long COVID—that uses mindfulness-based attention training, goal-management strategies, and fatigue-management techniques delivered through group and individual sessions with assigned practice. In contrast, the tDCS intervention is a neuromodulation approach in which a home-based device delivers low-intensity electrical stimulation to the dorsolateral prefrontal cortex while participants perform BrainHQ, aiming to enhance neuroplasticity and amplify the effects of cognitive training.

In the trial, lead author David S. Knopman, MD, a clinical neurologist at Mayo Clinic, and others, examined the effects of these interventions on 328 patients who met the criteria for Long COVID. To do so, investigators used change in the modified Everyday Cognition Scale 2 (ECog2) as the primary outcome. Patients in the study had efficacy assessments midway through the intervention at day 35 (±3) and at the end of the intervention (EOI) at day 70 (±3).

READ MORE: Decoding the Circuitry That Stabilizes Memory Maps in the Hippocampus

All told, none of the interventions showed efficacy, with no observed differences on the modified ECog2 for any of the prespecified primary comparisons. Overall, the adjusted differences in mean change for the primary contrasts were as follows: BrainHQ vs active comparator, 0.0 (95% CI, −0.2 to 0.2); PASC-CoRE + BrainHQ vs active comparator, 0.1 (95% CI, −0.1 to 0.3); tDCS-active + BrainHQ vs tDCS-sham + BrainHQ, 0.0 (95% CI, −0.2 to 0.2); and PASC-CoRE + BrainHQ vs BrainHQ alone, 0.1 (95% CI, −0.1 to 0.3).

In the trial, participant-reported outcomes and neuropsychological tests showed no meaningful differences between treatment arms. Across all groups, secondary measures—such as sleep, fatigue, cognition, mood, and several cognitive test scores—improved from baseline to the end of intervention. Subjective global ratings also reflected broad perceived benefit: 74% reported improved cognitive functioning and 66% reported better overall functioning, while only 7% felt somewhat or much worse.

In terms of safety, a total of 214 participants (76%) reported postexertional malaise at baseline, compared with 143 participants (56%) at the end of intervention, with some variation across treatment arms; none discontinued because of it. In the modified ITT population (n = 320), 11 participants experienced treatment-emergent serious adverse events, including one minor thermal injury under an electrode in a participant receiving active tDCS about one month after randomization.

The RECOVER-NEURO trial was strengthened by its randomized design, multiple evidence-based interventions, large national sample, comprehensive assessments, and high retention and adherence supported by a centralized, contactless telehealth framework. However, uncertainties surrounding the biology of cognitive long COVID, potential benefits from the active comparator or nonspecific trial effects, and the inclusion of many participants without objective cognitive impairment make it harder to detect differences between treatments. The study authors noted that future trials should examine whether patients with documented cognitive deficits respond differently and consider stratifying participants by objective cognitive function.

REFERENCE
1. Knopman DS, Koltai D, Laskowitz D, et al. Evaluation of Interventions for Cognitive Symptoms in Long COVID: A Randomized Clinical Trial. JAMA Neurol. Published online November 10, 2025. doi:10.1001/jamaneurol.2025.4415