Cognitive Improvement and Brain Changes after Real-Time Functional MRI Neurofeedback Training in Healthy Elderly and Prodromal Alzheimer’s Disease

Christian Hohenfeld, Nils Nellessen, Imis Dogan, Hanna Kuhn, Christine Müller, Federica Papa, Simon Ketteler, Rainer Goebel, Armin Heinecke, N. Jon Shah, Jörg B. Schulz, Martina Reske, Kathrin ReetzFrontiers in Neurology 2017 · DOI

Abstract

These findings suggest that cognitive decline, either related to prodromal AD or healthy ageing, could be counteracted using fMRI-based neurofeedback. Future research needs to determine the potential of this method as a treatment tool.

Introduction

Cognitive decline, particularly in visuospatial memory—the ability to navigate and remember locations—is a hallmark of both healthy aging and Alzheimer’s Disease (AD). The parahippocampal gyrus (PHG) serves as a critical neural hub for these functions but is often among the first regions to undergo atrophy in AD. While traditional cognitive interventions like "brain games" offer some benefits, this study explores a direct "top-down" intervention: real-time functional MRI (rtfMRI) neurofeedback. The central question was whether elderly brains, including those with prodromal AD, retain sufficient plasticity to master self-regulation of specific brain regions and if this mastery translates into cognitive recovery.

Methods

The study utilized a quasi-experimental longitudinal design with 30 participants: 16 healthy elderly (HC), 10 patients with prodromal AD (PA), and 4 healthy controls receiving sham feedback.

Experimental protocol showing the timeline from pre-test (T1) through neurofeedback training (T2-T4) to post-test (T5).

The protocol involved several key stages:

  1. Encoding: Participants memorized a 2km real-world footpath on a university campus.
  2. Neurofeedback Training: During MRI sessions, participants viewed a "thermometer" visual representing the BOLD signal from their left PHG. They were instructed to "upregulate" this signal by mentally retracing the learned path.
  3. Real-Time Analysis: Data was processed instantly using Turbo-BrainVoyager to provide immediate feedback to the participant.
  4. Structural and Connectivity Analysis: Voxel-Based Morphometry (VBM) tracked physical changes in grey matter, while Granger Causality Analysis (GCA) mapped directional influences between brain regions.
Brain slices highlighting the parahippocampal gyrus (PHG) regions targeted during the neurofeedback sessions.

Results and Key Insights

The intervention yielded significant cognitive and structural improvements across the experimental groups.

Cognitive and Structural Plasticity

Both healthy elderly and prodromal AD patients demonstrated significant gains in visuospatial memory (measured via the VVM test). Healthy subjects also showed improvements in working memory and overall MoCA scores. Notably, VBM analysis revealed increased grey matter volume in the right precuneus and right superior medial frontal gyrus. This suggests that even brains affected by early-stage AD maintain a degree of structural plasticity.

Neural Network Reorganization

The precuneus emerged as a "hidden driver" of the neural network during the tasks, showing increased connectivity and volume. This indicates its role as a support hub for mental imagery. Interestingly, in healthy subjects, the left PHG transitioned from a "driver" to a "receiver" of activation, suggesting a more efficient reorganization of the memory retrieval network.

The Importance of Specificity

The sham control group, which received feedback from unrelated brain regions, showed no cognitive improvement. This validates that the observed benefits were specifically due to the targeted neurofeedback rather than a general placebo effect or the mere effort of being in the scanner.

Discussion and Limitations

While the results are promising, several limitations must be addressed. The small sample size, particularly in the sham group ($n=4$), necessitates larger trials for clinical validation. Additionally, the lack of long-term follow-up leaves the durability of these changes unknown. AD patients also reported greater difficulty in controlling the feedback compared to healthy controls, suggesting a need for more accessible or simplified feedback interfaces. Finally, while rtfMRI is a powerful research tool, its high cost may limit widespread clinical application, prompting interest in adapting these protocols for EEG or fNIRS.

In conclusion, rtfMRI neurofeedback represents a potent potential treatment for age-related and prodromal AD cognitive decline by leveraging the brain's inherent plasticity to reorganize and strengthen critical memory networks.

Summary updated

Abstract

These findings suggest that cognitive decline, either related to prodromal AD or healthy ageing, could be counteracted using fMRI-based neurofeedback. Future research needs to determine the potential of this method as a treatment tool.
BibTeX
@inproceedings{hohenfeld2017cognitive,
  author = {Christian Hohenfeld and Nils Nellessen and Imis Dogan and Hanna Kuhn and Christine Müller and Federica Papa and Simon Ketteler and Rainer Goebel and Armin Heinecke and N. Jon Shah and Jörg B. Schulz and Martina Reske and Kathrin Reetz},
  title = {{Cognitive Improvement and Brain Changes after Real-Time Functional MRI Neurofeedback Training in Healthy Elderly and Prodromal Alzheimer’s Disease}},
  booktitle = {Frontiers in Neurology},
  year = {2017},
  doi = {10.3389/fneur.2017.00384},
}
Figure