Publications

This foundational study demonstrates that sub-scalp EEG contains the high-fidelity sensorimotor signals required for speech decoding. It validates the core premise of Fluent: that high-bandwidth data is observable without penetrating the skull.

The definitive safety trial for the bilateral subscalp implant. This study confirmed that the hardware platform is safe for long-term use (over 12 months) and provides signal clarity comparable to scalp EEG.

A 15-month prospective study confirming the reliability and signal stability of sub-scalp EEG in humans during everyday life. Provides critical evidence for the device's "set-and-forget" capability.

Detailed engineering specifications for the 'Set-And-Forget EEG (SAFE)' system. This paper outlines the custom amplifier and wireless transmission architecture optimized for chronic BCI applications.

A direct head-to-head comparison demonstrating that sub-scalp electrodes achieve signal bandwidth and signal-to-noise ratios comparable to invasive endovascular sensors (Stentrode), validating the "Goldilocks" approach.

Investigates the neural signatures of phonemes (the building blocks of speech) using intracranial data, providing the algorithmic groundwork for Fluent's speech decoding pipeline.

The first-in-human clinical trial of the Stentrode (Synchron). Our team members (Morokoff, John, Grayden) were instrumental in demonstrating that a minimally invasive BCI could safely restore digital freedom to paralyzed patients.

Demonstrates the power of continuous sub-scalp data for predictive algorithms. While focused on seizures, this establishes the device's capability to feed real-time, AI-driven forecasting models.

Validates the use of LSTM (Long Short-Term Memory) networks for real-time biological signal forecasting, a core component of Fluent's AI Scaffold architecture.

A defining review of the sub-scalp EEG landscape, outlining the clinical necessity and technological convergence that drives the shift from acute hospital monitoring to chronic, invisible home monitoring.

Proves that Deep Learning (CNNs) can decode motor intent (hand squeezing) from EEG signals, a critical step toward the complex motor decoding required for speech interfaces.

Identifies 'critical slowing down' as a universal warning signal in brain dynamics, providing a powerful biomarker for predicting state transitions (like seizures or intention) in chronic recordings.

A high-impact study analyzing long-term cycles in brain activity. This work demonstrated the team's capability to handle and analyze massive longitudinal neural datasets.

The foundational engineering study that established the "peg" electrode design. It proved that this specific form factor offers superior signal quality and mechanical stability compared to standard disk electrodes.

The groundbreaking paper that introduced the Stentrode. Our co-founders (Grayden, John) were key authors, proving the viability of recording brain signals from inside a blood vessel—a major leap in minimally invasive BCI.

A seminal paper on Attention Mechanisms in NLP. The hierarchical approach described here informs how Fluent's AI Scaffold prioritizes phonemes, words, and sentences during decoding.

Recognition of the original development grant led by Prof. Mark Cook (with Prof. Grayden and Dr. Freestone) as one of the top 10 translational research projects in Australia, launching the subscalp EEG platform.

The historic "First-in-Man" study for intracranial seizure forecasting. It set the benchmark for all future long-term BCI trials and established the team's capability to run complex implantable device trials.