Basal Ganglia as an fMRI Motor Neurofeedback Target in Parkinson’s Disease: the question, the experiment, and why it answers — a commentary on Baqapuri et al. (Applied Psychophysiology and Biofeedback, 2025)

1) The scientific question

The core question is: Can people with Parkinson’s disease voluntarily modulate basal ganglia activity—specifically the putamen—using real-time fMRI neurofeedback during motor imagery?
A second, practical question follows: Does directly targeting the putamen produce stronger putamen engagement than using a classical cortical target such as the supplementary motor area (SMA)?

Basal Ganglia as an fMRI Motor Neurofeedback Target in Parkinsons Disease

2) The experiment

The authors ran two rt-fMRI neurofeedback (NF) studies, both using a within-subject crossover comparison of two targets:

• Healthy cohort: 12 participants, one MRI session, with NF runs targeting putamen and SMA in separate runs. Participants were blinded to which ROI was targeted in each run.

• Parkinson cohort: 12 patients, three MRI sessions, approximately one week apart, with target order reversed across sessions.

Task: participants used motor imagery to upregulate the feedback “thermometer” during NF blocks compared with rest blocks.
Movement confound control: EMG (and in a subset of healthy participants, eye tracking) to verify that effects were not driven by overt movement or gaze differences.

Primary outcomes included ROI activation (NF > rest), potential learning effects across sessions, and differences between conditions (putamen targeted vs SMA targeted).

3) Why this experiment answers the question

It answers the question in three direct ways:

1. If the question is “can the putamen be modulated?”, the ROI contrast NF > rest tests whether putamen BOLD increases during NF-guided motor imagery.

2. If the question is “does direct targeting matter?”, the crossover design compares putamen activity when it is the explicit target versus when feedback comes from the SMA.

3. If the concern is “is this just movement?”, EMG (and eye tracking in some participants) tests the most likely confound: covert or overt muscular activation and visual tracking differences.

4) What the results indicate (functionally)

• Parkinson patients successfully recruited the putamen during NF-guided motor imagery (significant NF > rest effect).

• No clear learning effect emerged across sessions.

• Putamen activation did not differ meaningfully between “putamen-target” runs and “SMA-target” runs, suggesting strong network co-activation rather than strict target specificity.

• The exploratory whole-brain maps showed widespread cortical and subcortical motor-network engagement, consistent with motor imagery and neurofeedback as network-level processes.

5) BrainLatam reading — APUS (extended proprioception)

We read this study as an attempt to train APUS through the motor control circuit. Motor imagery is not just “thinking about movement”—it engages predictive sensorimotor organization. The fact that patients can modulate putamen activity suggests that APUS-related plasticity remains accessible even under dopaminergic impairment.

The finding that putamen activation was similar when targeting SMA implies an important operational point: the brain regulates circuits more than isolated nodes. Training a cortical “entry point” (SMA) can bring subcortical engagement along with it—useful if the intended intervention is network-level regulation rather than ROI purity.

6) BrainLatam reading — Tekoha (extended interoception)

Neurofeedback training also depends on internal state stability—sustained attention, fatigue tolerance, and consistent strategy deployment. The difficulty sustaining regulation over time and the absence of clear learning effects may reflect interoceptive variability across sessions, especially in Parkinson’s disease.

The involvement of regions such as the insula in exploratory analyses fits this view: self-regulation tasks often recruit interoceptive and autonomic control systems that support maintaining a stable engagement state.

7) Key limits that define the next experiment

• Small samples (12 + 12) reduce sensitivity to detect learning and target-specific differences.

• No sham feedback condition limits separation of “neurofeedback effect” from “motor imagery effect” (the authors note feasibility concerns in clinical cohorts).

• This is a feasibility study; clinical transfer (e.g., symptom change, real-world motor improvement) is not tested here.

• The lack of learning suggests future protocols may need more sessions, distributed practice, more structured strategy guidance, or more informative feedback (e.g., connectivity-based or multivariate pattern feedback rather than simple percent signal change).

8) BrainLatam translation to the organic world

BrainLatam translation to the organic world: this work supports the feasibility of non-invasive self-regulation of a subcortical motor hub relevant to Parkinson’s disease. Even without strong learning effects, the demonstration that patients can upregulate putamen activity through guided self-regulation indicates that the motor circuit retains trainable degrees of freedom. Future interventions may benefit from explicitly treating neurofeedback as circuit training, integrating cortical entry points and basal ganglia targets within longer and more precise protocols.

9) Open BrainLatam question

If the brain regulates circuits rather than single points, what is the best neurofeedback target in Parkinson’s disease:

• a single ROI (putamen),

• a cortical entry node (SMA), or

• a circuit marker (e.g., SMA–putamen–thalamus connectivity) that directly reflects APUS in action?

The body does not need belief to function.
It needs space, movement, and regulation.

Ref.:

‌Baqapuri, H. I., Terneusen, A., Luehrs, M., Peters, J., Kuijf, M., Goebel, R., Linden, D., Lozano, A. M., Mana, J., Jarraya, B., Loução, R., Kocher, M., Visser-Vandewalle, V., & Cukur, T. (2025). Basal ganglia as an fMRI motor neurofeedback target in Parkinson’s disease. Applied Psychophysiology and Biofeedback, 50(4), 635–653. https://doi.org/10.1007/s10484-025-09747-5