Months-long tracking of neuronal ensembles spanning multiple brain areas with Ultra-Flexible Tentacle Electrodes

Understanding the complex dynamics of neuronal ensembles across distributed brain networks is crucial for unraveling the neural mechanisms underlying cognition, memory, and behavior. Traditional neural recording techniques, however, often suffer from limitations in recording duration, spatial resolution, and the ability to track the same neurons over extended periods. Rigid electrode arrays can cause significant tissue damage, leading to inflammation, glial scarring, and a decline in signal quality over time. Flexible electrodes have shown promise in mitigating these issues, but challenges remain in achieving high channel counts, deep brain penetration, and long-term stability. This research addresses these limitations by introducing Ultra-Flexible Tentacle Electrodes (UFTEs), a novel technology designed for months-long, high-density recordings from multiple brain regions in freely moving animals. The unique design of UFTEs, combining mechanical and chemical tethering for insertion, allows for the simultaneous implantation of many independent recording fibers with minimal tissue disruption. The small footprint of UFTEs and their inherent flexibility promote long-term stability and reduce the likelihood of neuronal loss or displacement. This study aimed to evaluate the performance and biocompatibility of UFTEs, and to utilize them to investigate the dynamics of neuronal ensembles spanning multiple brain areas over several months. The importance of this research lies in its potential to significantly advance our understanding of long-term neural dynamics, enabling new insights into learning, memory, and the pathophysiology of brain disorders.

The field of chronic neural recording has seen significant advancements with the development of flexible electrode arrays, aiming to overcome the limitations of rigid probes that cause tissue damage and recording instability. Studies have demonstrated the importance of long-term recordings for understanding neuronal plasticity and the dynamics of neural circuits. For example, studies using silicon probes, such as NeuroPixels, have enabled long-term recordings of neuronal activity, but these often involve significant tissue disruption and challenges with drift correction. Other flexible electrode designs, such as neural meshes, have addressed some of these challenges, but limitations persist in achieving both high channel counts and deep brain penetration. Furthermore, methods for reliably tracking individual neurons across multiple recording sessions remain a significant challenge. The current literature highlights the need for a technology capable of achieving long-term stable recordings with high signal-to-noise ratios and minimal tissue damage, enabling the investigation of multi-regional neuronal interactions and the dynamics of neuronal ensembles over extended periods. This research seeks to fill this gap by providing a new technology with superior characteristics compared to existing methods.

The study involved the fabrication, characterization, and in vivo implantation of UFTEs in rodents. UFTEs were fabricated using a microfabrication process involving polyimide layers, gold electrodes, and PEDOT:PSS coating to reduce impedance. The electrodes consisted of numerous independently moving ultra-flexible polyimide fibers bundled together for implantation and mechanically tethered to a tungsten shuttle for insertion. A biodegradable glue held the fibers together during insertion, after which the shuttle was removed, leaving the independent fibers in the brain. A 512-channel wireless logger was used for recording local field potentials (LFPs) and spiking activity. The implantation procedure involved simultaneous insertion of multiple bundles into different brain regions (hippocampus, retrosplenial cortex, medial prefrontal cortex) using a stereotactic arm and custom titanium headstage to minimize damage and ensure stability. Electrophysiological recordings were performed in freely moving rats over several months. Biocompatibility was assessed through immunostaining for markers of tissue damage and inflammation. Spike sorting was performed using a semi-automatic clustering method, followed by manual correction. Neuronal ensembles were identified as groups of neurons displaying repeated co-activation within a narrow time window (25ms). Ensemble lifetimes and activation strength during sharp-wave ripples (SWRs) were analyzed to characterize ensemble dynamics. Statistical analysis was performed using Student’s t-test and Wilcoxon rank-sum test.

The UFTEs demonstrated exceptionally high signal-to-noise ratios (SNRs) and unit yields, surpassing state-of-the-art flexible electrode arrays. Mean SNRs were 1.5 times higher, with some reaching 89. Average cortical unit yields were 1.75/channel. Importantly, immunostaining revealed no detectable chronic tissue damage after several months of implantation. UFTEs enabled the successful tracking of the same neurons for at least 10 months (the longest duration tested). Simultaneous recordings from the hippocampus, retrosplenial cortex, and medial prefrontal cortex revealed the existence of both intra-areal and inter-areal neuronal ensembles. The analysis revealed two distinct classes of ensembles: those strongly tuned to SWRs, which displayed shorter lifetimes and were predominantly hippocampal; and those weakly tuned to SWRs, which had longer lifetimes (up to months) and often included neurons from both the hippocampus and cortex. The long-lived inter-areal ensembles exhibited periods of inactivity before re-emerging, suggesting a potential role in long-term memory processes. The median ensemble lifetime was significantly shorter than the duration for which individual neurons were trackable (4 days vs 24 days). Analysis of ensemble activation strength and SWR tuning further supported the distinction between short-lived, SWR-tuned hippocampal ensembles and long-lived, weakly SWR-tuned inter-areal ensembles.

The results demonstrate the significant advantages of UFTEs for long-term, high-density neural recordings. The high SNRs and unit yields, combined with the minimal tissue damage and long-term stability, offer substantial improvements over existing technologies. The ability to track the same neurons and ensembles over months allows for unprecedented insights into the dynamics of neural circuits involved in learning and memory. The identification of two distinct classes of ensembles, one characterized by short lifetimes and strong SWR tuning, and the other by longer lifetimes and weak SWR tuning, suggests the existence of multiple mechanisms for information processing and storage in the brain. The long-lived inter-areal ensembles may play a crucial role in long-term memory consolidation and retrieval. Future research could explore the specific functional roles of these ensembles and their involvement in different cognitive processes. The development of UFTEs opens new avenues for investigating long-term neural plasticity and the pathophysiology of neurological and psychiatric disorders.

This study successfully demonstrates the feasibility and effectiveness of UFTEs for long-term, high-density neural recordings. UFTEs provide significant advantages over existing technologies in terms of signal quality, stability, and biocompatibility, enabling the tracking of individual neurons and neuronal ensembles across multiple brain regions over extended periods. The findings highlight the existence of distinct classes of neuronal ensembles with varying lifetimes and functional properties, providing new insights into brain dynamics and information processing. Future research should explore the clinical applications of UFTEs in brain-computer interfaces and the investigation of brain disorders.

While UFTEs offer significant advantages, some limitations exist. The current study was conducted in rodents, and further research is needed to assess the effectiveness and scalability of UFTEs in larger animals and humans. The analysis focused on specific brain regions and cell types; future investigations could expand the scope to include a broader range of brain areas and neuronal populations. The methods used for ensemble identification may not capture all forms of neural coordination; alternative approaches could be explored to enhance the comprehensiveness of ensemble analysis. Also, the long-term stability of the recording, although impressive, still shows some dynamism in the single unit detection, which might be caused by neuronal movement or death.