Near-Infrared Spectroscopy – Electroencephalography-Based Brain-State-Dependent Electrotherapy: A Computational Approach Based on Excitation–Inhibition Balance Hypothesis

Snigdha Dagar, Shubhajit Roy Chowdhury, Raju S. Bapi, Anirban Dutta, D. PomeroyFrontiers in Neurology 2016 · DOI

Abstract

Stroke is the leading cause of severe chronic disability and the second cause of death worldwide. Post-stroke chronic disability may be ameliorated with early neurorehabilitation where non-invasive brain stimulation (NIBS) techniques can be used as an adjuvant treatment. However, the heterogeneity in the lesioned brain requires individualized NIBS intervention. This paper proposes Brain State Dependent Electrotherapy (BSDE), leveraging portable electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) for real-time monitoring. We propose a computational approach based on the excitation-inhibition (E-I) balance hypothesis to quantify post-stroke brain states and optimize NIBS using individual forward models and neural mass modeling.

Introduction: The Challenge of Individualized Stroke Recovery

Stroke remains a primary driver of long-term disability, frequently impairing motor and cognitive functions. While non-invasive brain stimulation (NIBS), particularly Transcranial Direct Current Stimulation (tDCS), has shown potential in "priming" the brain for rehabilitation, clinical outcomes are often inconsistent. This inconsistency stems from a "one-size-fits-all" approach that fails to account for the unique neurophysiological landscape of each patient.

Stroke lesions create a highly heterogeneous environment where the local balance of excitation (mediated by Glutamate) and inhibition (mediated by GABA) is disrupted. A stimulation protocol beneficial for one patient might be ineffective or even detrimental for another. Consequently, there is a critical need for systems capable of monitoring an individual’s brain state in real-time and adjusting the electrotherapy "dose" to restore physiological E-I balance.

Methods: A Multi-Modal Computational Pipeline

The proposed BSDE framework integrates real-time neuroimaging with advanced computational modeling to create a personalized, closed-loop rehabilitation system.

Joint Imaging and Sensor Montage

The system utilizes a dual-modality approach combining Electroencephalography (EEG) to capture fast electrical oscillations and functional Near-Infrared Spectroscopy (fNIRS) to monitor slower hemodynamic changes.

Experimental setup and sensor montage showing the integration of HD-tDCS, fNIRS, and EEG sensors.

Neural Mass Modeling (NMM)

At the heart of the approach is the Neural Mass Model. Rather than modeling individual neurons, NMMs simulate the collective behavior of large populations, such as excitatory pyramidal cells and inhibitory interneurons. The authors employ a thalamocortical NMM to bridge the gap between EEG electrical signals and fNIRS metabolic data via neurovascular coupling.

Adaptive Control Loop

The system employs a Kalman filter to estimate the hidden states of the neurovascular unit (NVU) from imaging data. An artificial neural network then maps these states to optimal tDCS parameters (amplitude and duration), targeting a specific "operating point" of E-I balance.

Key Insights: The E-I Balance and Forward Modeling

The core hypothesis is that post-stroke dysfunction is essentially a "tuning" problem. By quantifying the E-I ratio, clinicians can move toward objective, biophysical targets.

Computational forward model illustrating the predicted current flow or activation patterns in the brain.

A critical component is the use of Personalized Forward Models. These models predict how electrical current flows through a specific patient's lesioned brain, allowing for the optimization of electrode placement and current density to accurately hit the target region. Furthermore, the framework views tDCS as a neuromodulator that assists the brain's homeostatic mechanisms in returning to a stable, healthy regime.

Results and Validation

While primarily a theoretical framework, the paper provides validation through simulations and pilot data:

  • Spectral Matching: The NMM successfully replicated empirical EEG findings, showing prominent power spectral peaks in the Alpha (10–12 Hz) and Beta (15–20 Hz) bands.
  • Sensitivity Analysis: Simulations confirmed that Alpha band amplitude is highly sensitive to changes in inhibitory synaptic strength, validating its use as a marker for E-I balance.
  • Polarity Effects: The model demonstrated that anodal stimulation increases excitatory firing rates and desynchronizes the mu-rhythm, while cathodal stimulation has the opposite effect.

Limitations and Future Directions

Despite its promise, several hurdles remain. Large-scale clinical trials are necessary to prove that this closed-loop approach outperforms standard tDCS in functional recovery. Additionally, real-time execution of complex NMMs requires significant processing power, which may challenge the development of portable bedside devices. Future work could incorporate structural reserve data (e.g., DTI scans) to further refine individualization and explore how acute E-I shifts translate into permanent long-term potentiation.

Summary updated

Abstract

Stroke is the leading cause of severe chronic disability and the second cause of death worldwide with 15 million new cases and 50 million stroke survivors. The poststroke chronic disability may be ameliorated with early neuro rehabilitation where non-invasive brain stimulation (NIBS) techniques can be used as an adjuvant treatment to hasten the effects. However, the heterogeneity in the lesioned brain will require individualized NIBS intervention where innovative neuroimaging technologies of portable electroencephalography (EEG) and functional-near-infrared spectroscopy (fNIRS) can be leveraged for Brain State Dependent Electrotherapy (BSDE). In this hypothesis and theory article, we propose a computational approach based on excitation-inhibition (E-I) balance hypothesis to objectively quantify the poststroke individual brain state using online fNIRS-EEG joint imaging. One of the key events that occurs following Stroke is the imbalance in local E-I (that is the ratio of Glutamate/GABA), which may be targeted with NIBS using a computational pipeline that includes individual "forward models" to predict current flow patterns through the lesioned brain or brain target region. The current flow will polarize the neurons, which can be captured with E-I-based brain models. Furthermore, E-I balance hypothesis can be used to find the consequences of cellular polarization on neuronal information processing, which can then be implicated in changes in function. We first review the evidence that shows how this local imbalance between E-I leading to functional dysfunction can be restored in targeted sites with NIBS (motor cortex and somatosensory cortex) resulting in large-scale plastic reorganization over the cortex, and probably facilitating recovery of functions. Second, we show evidence how BSDE based on E-I balance hypothesis may target a specific brain site or network as an adjuvant treatment. Hence, computational neural mass model-based integration of neurostimulation with online neuroimaging systems may provide less ambiguous, robust optimization of NIBS, and its application in neurological conditions and disorders across individual patients.
BibTeX
@inproceedings{dagar2016near,
  author = {Snigdha Dagar and Shubhajit Roy Chowdhury and Raju S. Bapi and Anirban Dutta and D. Pomeroy},
  title = {{Near-Infrared Spectroscopy – Electroencephalography-Based Brain-State-Dependent Electrotherapy: A Computational Approach Based on Excitation–Inhibition Balance Hypothesis}},
  booktitle = {Frontiers in Neurology},
  year = {2016},
  doi = {10.3389/fneur.2016.00123},
}
Figure