A functional view of the tumour–brain interface

Gliomas do not exist apart from the brain. They grow within electrically active tissue, interact with resident cells and neural circuits, and alter the environment they infiltrate. Cancer neuroscience has established that nervous-system activity and cancer biology can influence one another. Our work focuses on how this interaction can be observed in living tissue with enough spatial, temporal and physical detail to be useful.

We approach the problem from neural engineering and experimental neurophysiology. Rather than assuming that one signal will define a tumour margin or tissue state, we test how complementary measurements behave across tissue regions and how they relate to established anatomical, histological and molecular descriptions.

1. Electrophysiology across heterogeneous tissue

We investigate electrical activity across apparently healthy brain, infiltrated tissue, tumour boundary and tumour core. We ask:

• Which features of neuronal and tissue-level activity change across these regions?
• Which observations hold across preparations, models and measurement conditions?
• How can local electrophysiology be read alongside tumour biology rather than in isolation?
• What does a functional boundary mean in tissue whose anatomy, cellular make-up and activity change continuously?

This is an active research programme. The lab studies candidate signals and measurement strategies; it does not assume that electrophysiology alone can identify tumour tissue or define a safe surgical boundary.

2. Multimodal descriptions of tissue state

Electrical activity is one part of a wider tissue environment. Ionic composition, metabolites, impedance and molecular vibrations each describe different physical processes and operate at different scales. We combine electrophysiology with electrochemistry, impedance measurement, Raman and other optical spectroscopy, spatial biology and computational analysis, looking for measurements that are complementary, interpretable and stable enough to test a defined biological or translational question. In practice, we aim to:

• measure different domains as close together in space and time as the experiment permits;
• keep track of where each signal came from and how uncertain it is;
• compare candidate readouts with relevant biological reference measurements;
• test whether findings generalise, rather than optimising within a single dataset; and
• treat a multimodal model as useful only when its added value can be demonstrated.

3. Interfaces that preserve the system being studied

Neural interfaces can change the tissue they are meant to measure. Probe size, materials, geometry, insertion and electrochemical properties all affect the quality and interpretation of a recording.

The lab’s technical lineage includes published work on jULIEs, micrometre-scale, low-impedance electrodes for extracellular recording and stimulation, and CHIME, a strategy for coupling three-dimensional microwire arrays to CMOS electronics for scalable neural recording, demonstrated at large scale in related collaborative work. These studies were carried out with multidisciplinary collaborators in experimental neuroscience settings. We build on their principles when choosing or developing interfaces for tumour-bearing brain: obtain useful signals, limit perturbation, understand the electrode–tissue relationship, and design for the biological question rather than for channel count alone.

4. Translation as an evidence pathway

The translational programme asks how experimental measurements could eventually become reliable and useful in a surgical or clinical research setting. That takes more than a working prototype: analytical validation, preclinical evidence, usability, workflow integration, quality management, regulatory planning, clinical-study design, patient and public involvement, and comparison with current practice.

Longer-term directions include functional mapping of tumour and surrounding brain, longitudinal measurement of tissue and circuit change, and experimental therapeutic neurotechnology informed by tumour–brain biology. These are directions under investigation, not statements of established clinical benefit.

5. Neuromorphic computation and non-von Neumann hardware

Neural and physicochemical measurements can produce varied, high-rate data under tight limits on latency, power and size. As interfaces gain more channels, moving and processing all the raw data can come to dominate power, bandwidth and experimental complexity. We are interested in neuromorphic computation, in-memory processing, event-driven sensing and other non-von Neumann hardware as possible ways to:

• process electrophysiological and multimodal signals close to the sensor;
• reduce data movement and energy use;
• respond to sparse or asynchronous biological events;
• support adaptive and closed-loop experiments; and
• explore hardware–algorithm co-design for biomedical measurement, including neuromorphic brain–computer interfaces.

This direction builds on our published neural-interface work. The scientific question is where these architectures offer a measurable advantage over conventional computing for a defined cancer-neurotechnology task. We are seeking funding and partners to develop it into an integrated research platform.

Why this matters

For people affected by brain tumours, the central problem is not a shortage of bold claims. It is the need for measurements and interventions that are demonstrably informative, safe and relevant to decisions that matter. Our contribution is to examine the functional and physicochemical information present in living tissue, develop ways to access it, and test, step by step, whether it can complement current biological and clinical knowledge.

Background reading

• Mancusi R, Monje M. The neuroscience of cancer. Nature. 2023.
• Winkler F, et al. Cancer neuroscience: state of the field, emerging directions. Cell. 2023.
• Venkatesh HS, et al. Electrical and synaptic integration of glioma into neural circuits. Nature. 2019.

These are independent studies by other groups that inform our questions; they are not the lab’s own work.