Outline of potential overall thesis structure

Section 1: Intro (short - focus on targeted reviews of relevant topics)
- Broad: prostheses, functional restoration
- Computational methods
- Modeling for patient-specific stimulation
- Spatiotemporally precise stimulation & control of neural systems

Section 2: Theory (short - focus on developing theory and demonstrating rationale; formal statement of overall system)
- Develop theory and formalisms for project (using 2D worked examples for demonstration)
- Solving PDEs on patient specific domains (patient-specific modeling)
- Approximating neural responses to extracellular fields (activation, axon models)
- Derivative-free optimisation with physiologically relevant objective functions (pattern search, genetic algorithms, RL; get ideal parameters)
- Neural control systems (once optimal parameters obtained, timing of stimulation and adaptive stimulation)
(-> overall system for getting optimal stimulation parameters and control algorithms for modulating specific neural systems)

Section 3: Implementation (longer - technical details of actual methods and results)
- nnrd-model: create models from images and solve V (including interferential)
- nrrd-biophysics: rotated Hessian for activation & interpolating onto axons at target
- nrrd-optimise: finding optimal parameters for stimulation
- nrrd-print: create models for validation
- control signals?: cuff (adaptive), phrenic (detection for triggering), BCI (DL regression)

Section 4: Validation (longer - detailed validation protocols and results)
- Physical models: construction and testing
- In human models: paraesthesias, pick up on trial leads, response with phrenic, median
- Cuff, phrenic phase detection testing, BCI analysis (3D CNN)

Section 5: Application (longer - detailed protocols for application and results)
- Noninvasive trialing
- Median nerve
- Stellate
- Phrenic nerve (targeting & control)
- Control algorithms: cuff, BCI, etc. (apply and use to determine rate of stimulation, etc.)
etc.

Section 6: Discussion / conclusions (short - summarise main points and indicate future potential impact)
- nrrdosurgery: full interface for integrated system for creating patient-specific models (validated)
- "Computational neurosurgery": paradigm for patient-specific modeling for optimal interventions
- Further trials
- Applications to other areas
- Move on to trials in SCI, etc.
- Full integration with control: e.g. use goniometric measures of limb position to determine stimulation pattern, etc.
	- ?build proof of principle system to drive noninvasive array of electrodes (interferential) based on control signal (e.g. goniometers, other wearables)

Appendices (short - outline notes for potential short course, etc.)
- Neuroanatomy: gross anatomy of spinal cord and peripheral nerves; microanatomy of nerve fibres
- Neurophysiology & biophysics: membrane potentials, excitable cells, neuron models; HH equations, multicompartment, cable; MRG, etc.
- Neural electrodynamics: electric fields, activation functions
- Mathematical modeling: linear algenra, calculus (incl vector, Jacobian, Hessian), PDEs, FEM
- Optimisation: standard, DFO, genetic, RL
- Control: control theory & applications
