Sources and References

Primary texts — Reis, Nordin, and Frontera, DeLisa’s Physical Medicine and Rehabilitation: Principles and Practice, 6th ed. (Wolters Kluwer). Horch and Dhillon (eds.), Neuroprosthetics: Theory and Practice, 2nd ed. (World Scientific).

Supplementary texts — Kandel et al., Principles of Neural Science, 6th ed. (McGraw-Hill). Winter, Biomechanics and Motor Control of Human Movement, 4th ed. (Wiley). Dhillon and Horch, Neuroprosthetics (World Scientific).

Online resources — MIT OCW HST.583 Functional Magnetic Resonance Imaging and 9.01 Neuroscience and Behavior. IEEE EMBS open lectures on neural engineering. NIH NINDS open resources on neurological disorders. ISO 14155 clinical investigation of medical devices. WHO International Classification of Functioning, Disability and Health.

Chapter 1: The Field of Neural and Rehabilitation Engineering

1.1 Scope

Neural engineering applies engineering methods to the nervous system: recording, stimulation, modulation, and neural prosthetics. Rehabilitation engineering overlaps strongly, extending to musculoskeletal dysfunction, assistive technology, and restoration of daily-living function. The two fields share patients, devices, regulatory pathways, and clinical-research methods.

1.2 Disability and Function

The WHO International Classification of Functioning, Disability, and Health (ICF) frames disability as interaction between body function/structure, activity, participation, and environment. Engineering interventions can target any level. The ICF replaces deficit-focused language with function-focused language, aligning engineering objectives with patient-meaningful outcomes.

Chapter 2: Neurological and Musculoskeletal Injury

2.1 Stroke

Ischemic and hemorrhagic stroke disrupt blood supply, producing focal neurological deficits whose severity depends on location and collateral circulation. Rehabilitation addresses motor, sensory, cognitive, speech, and swallowing deficits. Constraint-induced movement therapy, robotic-assisted training, and neuromuscular electrical stimulation each target motor recovery through different mechanisms of neuroplasticity.

2.2 Spinal Cord Injury

Complete and incomplete SCI produce loss of motor, sensory, autonomic, and bladder/bowel function below the lesion. ASIA Impairment Scale grades severity. Engineering responses span assistive technology (powered wheelchairs, BCIs), activity-based rehabilitation (body-weight-supported treadmill, epidural stimulation), and implanted neuromodulation.

2.3 Peripheral Nerve and Musculoskeletal

Peripheral nerve injury ranges from transient conduction block to complete axonal disruption (Seddon’s classification: neurapraxia, axonotmesis, neurotmesis). Musculoskeletal injuries include fractures, tendon and ligament damage, and chronic conditions (osteoarthritis). Orthoses, prostheses, targeted muscle reinnervation, and graded exercise constitute the therapeutic spectrum.

2.4 Neurodegenerative Disease

Parkinson’s, Huntington’s, ALS, and multiple sclerosis each produce characteristic deficits amenable to engineered intervention. Deep brain stimulation for Parkinson’s is a mature neural-engineering intervention; nascent approaches include closed-loop adaptive DBS and focused-ultrasound neuromodulation.

Chapter 3: Neural Interfaces

3.1 Recording

Recording electrodes span scalp EEG (µV-scale, poor spatial resolution, non-invasive), ECoG (mV-scale, sub-centimeter resolution, cortical surface), intracortical microelectrode arrays (single-unit resolution, most invasive). Neural signal quality depends on electrode impedance, interface stability, and biological response. Glial encapsulation degrades chronic recording — a central challenge in neural engineering.

3.2 Stimulation

Stimulation delivers charge-balanced biphasic pulses. Safe stimulation respects the Shannon criterion

with charge per pulse \( Q \) and electrode area \( A \). Exceeding this limit drives irreversible electrochemistry and tissue damage. Waveform shape (monophasic, biphasic, asymmetric), polarity, and frequency tune neural selectivity.

3.3 Signal Processing

Bandpass filtering, spike sorting, and feature extraction turn raw electrode signals into neural commands. Decoders — linear regression, Kalman filter, recurrent neural network — translate neural features into control outputs. Calibration across time is essential: recording non-stationarity requires adaptive decoders to maintain performance over days and weeks.

Chapter 4: Rehabilitation Engineering

4.1 Biomechanics and Motion Analysis

Marker-based motion capture, inertial measurement units, and electromyography quantify movement and muscle activity during functional tasks. Joint kinematics are computed by inverse kinematics; joint kinetics by inverse dynamics. Gait analysis in walking, for instance, identifies deviations such as reduced push-off, circumduction, or asymmetries that guide intervention.

4.2 Assistive Technology

Ranges from low-tech (canes, splints) to high-tech (powered exoskeletons, advanced prostheses, eye-gaze interfaces). Selection depends on user capability, environment, task, and cost. The ICF framework — body function, activity, participation — keeps attention on outcomes that matter to users, not just device specifications.

4.3 Functional Electrical Stimulation

FES uses surface or implanted electrodes to produce muscle contraction in paralyzed limbs for standing, cycling, grasp, and respiration. Device design respects fatigue (non-physiological recruitment order), selectivity (electrode configuration), and control (open-loop stimulation patterns or closed-loop with sensor feedback).

Chapter 5: Rehabilitation Robotics

5.1 Goals and Designs

Rehabilitation robots deliver repetitive, task-specific training at dose levels unachievable by therapists alone. Design archetypes include end-effector robots (handle at terminal), exoskeletons (matched anatomy), and ankle/wrist/elbow-specific devices. Degrees of freedom, actuation bandwidth, and force range are chosen for the targeted function.

5.2 Impedance Control

Robots that physically interact with humans require compliant behaviour. Impedance control shapes the apparent mass, damping, and stiffness seen at the interaction port:

Assist-as-needed paradigms reduce assistance with user performance, maximizing voluntary effort and neuroplastic driving.

5.3 Virtual Reality and Gamification

VR environments contextualize rehabilitation tasks and enhance engagement. Real-time feedback, adaptive difficulty, and measured performance support motor learning principles — specificity, intensity, salience. Evidence for added clinical benefit over conventional therapy is mixed; better-designed trials continue to accumulate.

Chapter 6: Clinical Research and Translation

6.1 Study Design

Rehabilitation trials face challenges: heterogeneous participants, blinding difficulty, placebo analogues, and meaningful outcome selection. Designs span feasibility (small N, safety and usability), efficacy (randomized controlled, explanatory), and effectiveness (pragmatic, real-world). Cluster randomization suits interventions delivered at clinic level.

6.2 Regulatory Considerations

Neural devices typically fall in higher risk classes (Class III U.S.; Class IV Canada for implantable active) due to invasiveness and criticality. Rehabilitation robots often qualify as Class II. Pathways — 510(k) for substantial equivalence, De Novo, PMA, HDE for humanitarian device exemption — are chosen based on precedent and evidence base. IDE submissions authorize investigational use.

6.3 Ethics

Vulnerable populations (acute injury, cognitive impairment) require careful consent processes. Surrogate decision-makers, ongoing capacity assessment, and freedom to withdraw without loss of standard care are non-negotiable. Post-trial access to investigational devices is an ethical obligation increasingly codified; abandoning a participant dependent on a neural interface is a recognized harm.

6.4 Design Framework

A synthesis framework: (1) define the target population and function clearly in ICF terms; (2) identify mechanism — restore, replace, augment, modulate; (3) specify signals, actuators, and coupling to biology; (4) design around safety constraints and long-term stability; (5) plan regulatory and clinical evidence aligned with risk class and mechanism novelty; (6) plan deployment ethics, including access and post-trial support. This framework is a practical ordering principle for rehabilitation-device design.