IntegraBrain: A multi-modal neural interface for the detection and suppression of focal epilepsy

Summary

Files contain the sum total of the research data presented within the thesis, 'IntegraBrain – A Multi-modal Neural Interface for the Detection and Suppression of Focal Epilepsy', by Spencer R Moore. Performed from 02/2020 till 07/2024 under the supervision of Prof Ivan Minev. This thesis can be sourced on the White Rose ethesis depository. Further details and directory can be found in the .zip files README.txt found at its first layer.

Research undertaken during this PhD was under the umbrella of the Integrated Implant Technology for Multi-modal Brain Interfaces ERC Starting Grant. The PhD's goal was to develop a soft, multi-modal neural interface that could record neural signals via electrocorticography and suppress epileptic seizures using focal cooling. Work produced from it covered 3D Bioprinting with silicones, In Silico thermal modelling, embedded firmware programming, electronic hardware design, In Vitro thermal testing and In Vivo studies.

The research conducted was done so in collaboration with Dr Thomas Paterson, Naomi King, Dr Jason Berwick, Dr Clare Howarth, Nick Hagis, Dr Arua da Silva, and Dr Shangcheng Chen.

Structure

The data contained within the .zip file is split into 5 folder:

1. 3D Printed Parts
2. Experimental Data
3. Hardware Schematics
4. Interface 3D Discovery Print Files
5. Programmes

A README file is located at the top folder level to provide data licensing info and further description of the folder structure. All raw data files contain column headers to describe what the data is and its units, and is also supplied in open formats (.txt, .csv, .stl, .svg) so they can be opened in the preferred processing application.

There are some unavoidable proprietary formats that exist and require licensed software to access/run.

• '.mph' files contain the simulation models to be run in COMSOL Multiphysics v5.5 using the Heat Transfer module and the Computer Aided Design Import Module. No known alternate software package exists to open and run the simulation model these files describe.
• '.vi' and '.lvproj' files are LabVIEW programmes written in National Instruments' proprietary programming language, G. A licensed version of LabVIEW 2023 or newer is required to run these programmes. No known software exists that is able to open and run LabVIEW VIs.
• '.mlx' or MATLAB Live Script files require MATLAB r2023 or newer to run for data processing. These MATLAB scripts can be opened independently of MATLAB using a text editor (e.g. notepad++) to review the methods used.
• '.bcd' are project files for regenHU's BioCAD (v1.1 - 17) software that is used to generate G-code instructions for the regenHU 3D Discovery r5 bioprinter. This software is now considered deprecated (as of 2021) due to that generation of bioprinter becoming obsolete. Both the print path design files (.svg) and outputted G-code (.iso) files are supplied in it stead. The G-code files can be opened with a text editor (e.g notepad++).
• '.uvprojx' the project configuration file type for the Keil uVision5 (5.37.0.0) integrated development environment (IDE). A community version of the MDK-ARM v6 (that contains the most up to date uVision build) can be sourced for free from the ARM Keil website. As the firmware was written in C99 and the source files (.c, .h) remain separated from the project, other embedded toolchains can be used with the armclang compiler to generate the firmware binaries.

Thesis Abstract

Focal cooling has been demonstrated as a promising treatment strategy for patients with medically intractable epilepsy. Cooling actuation is achieved via an invasive interface positioned in direct contact with neural tissue of the epileptic foci. Seizure suppression by focal cooling has been demonstrated extensively. However, pre-clinical proof-of-concept systems that have been produce thus far are too bulky and mechanically stiff. Long-term implantation of these devices would risk inducing significant compression injury and localised glial scaring over time.

In this thesis presents the efforts to assemble a miniaturised, multimodal interface that is capable of simultaneously performing electrocorticography (ECoG) recording, temperature monitoring and focal cooling. Comprising of a solid-state thermoelectric element, a microfluidic heat management system, integrated thermocouples, ECoG electrodes and supporting electronic circuitry, a systemic implementation approach was used to produce a tightly integrated system.

The Interface consists of discrete electronic components embedded in a mechanically soft matrix with the dimensions 9.5 x 9.5 x 2.5mm. This is realised through the leveraging of direct ink writing (3D printing) biocompatible silicones and electrically conductive Graphite-PDMS composites. Integration of composite electrodes allows for ECoG sensing whilst maintain device softness.

In Vitro interface systems verification was conducted using hydrogel-based neural tissue thermal models. At the interface-model tissue boundary, rapid cooling rates of ~3°Cs-1 were achieved during 14°C ΔT cooling runs. System power consumption is recorded to peak at 7.84W on cooling start and 2.6W during steady state of a 15°C ΔT cooling run. System maturation presented the opportunity to conduct acute In Vivo animal studies using the 4-AP seizure model. Initial data collected demonstrate reliable cortical tissue cooling down to 18°C. Comparison of nominal seizure recordings to those during active cooling indicate successful seizure suppression with seizure activity suppression increasing as cooling temperature decreases. A ~67% reduction in broadband seizure magnitude (0-30Hz) was recorded when cooling to 18°C.

Work presented here represents an evolution in focal cooling interface design beyond the existing solutions and a step towards a clinically viable device for patients.