Integration of MEG functional localization information in MRI studies for implantable device surgical planning and navigation

Abstract

Use of MEG functional localization for surgical interventions faces two challenges: different anatomical referential are used and both the implanted device and the surgical procedure have significant geometric constraints that have to be taken into account before generating a surgical planning. Commercially available neuro-navigation devices, while offering proprietary application programming interfaces (API) for multimodality data fusion lack MEG data handling, relying on sole DICOM import capabilities. We present a new method to introduce MEG functional localization data and implantable device characteristics as well as surgical parameters directly in DICOM compliant format with accuracy and safety checks to allow for surgical procedures guidance. Our method is based on transfer functions and a toolbox that allow: multi-referential data and geometrical constraints representation with integrity check; visual representation and generation of virtual MRI slices with space sampling error minimization and control; device constrained simulated ECoG with algorithm driven iterative electrode placement optimization. Clinical trial was approved by the ethics committee. We applied these methods to MEG recordings of imaginary motor tasks in 7 healthy subjects for BCI purposes. The BCI implant as developed at CLINATEC ® , consists of a 64 electrodes matrix. From simulated ECoG recordings, optimal electrode placement can be automatically computed. Implantation constraints were fulfilled: a maximal cortex-implant distance of 0.5 mm, a bone thickness of more than 6 mm. The graphical information of the optimal drilling axis and its 50 mm diameter as well as the electrode grid on the cortex allowed for a visual check of the feasibility of the intervention on the MRI. DICOM compliant slices were successfully transferred in a commercial neuro-navigation system.