Michelangelo Final Report - Alexander Camuto, Shankho Chaudhuri, Dexter Gajjar-Reid, Umar Hossain - BE3 HHCARD 2016

MichelangeloA THERAPEUTIC INTERFACE PUTTING PATIENTS IN

TOUCH WITH THEIR REHABILITATION

11. INTRODUCTION

2. CONCEPT GENERATION 2

3. DESIGN ANALYSIS

4. FUTURE WORK 5

3

5. CONCLUSION 5

6. REFERENCES 6

7. APPENDICES 7

AbstractMichelangelo is an ergonomically designed rehabilitative device for patients suffering from Peripheral Nerve Injury (PNI) and Complex Regional Pain Syndrome (CRPS) in their hands. It is intended to be used both in early stage rehabilitation starting in the six weeks following surgery, and to last through the entire rehabilitation journey. It has been created both for a therapist-guided clinical setting and for independent use at home. The model shown is a proof of concept prototype designed to demonstrate the potential of the device.

This report will provide the clinical context for Michelangelo, illustrate the design process (from initial ideas to final prototypes), include a comprehensive analysis of features, demonstrate how patients will use the device, and evaluate the device, identifying scope for improvements in the future.

CONTENTSALEXANDER CAMUTOSHANKHO CHAUDHURIDEXTER GAJJAR-REIDUMAR HOSSAIN

BE3-HHCARDHUMAN CENTRED DESIGN OF REHABILITATIVE AND ASSISTIVE DEVICES

DEPARTMENT OF BIOENGINEERINGIMPERIAL COLLEGE LONDON

SPECIAL THANKS TO:

DONNA KENNEDYClinical Specialist in Hand Therapy at Imperial College Healthcare NHS Trust

PAGE

INTRODUCTION1A series of lectures delivered by clinical professionals and a set of project briefs set by the Department of Bioengineering, Imperial College London, dictated the initial starting point for

this project: improving hand rehabilitation following surgery, specifically for patients suffering from Peripheral Nerve Injury (PNI) and Complex Regional Pain Syndrome (CRPS).

Throughout the project, from origin to final presentation, our priority was ensuring that any proposed solution was clinically relevant and beneficial. In

light of this, significant literature research was conducted to understand that needs of patients and users. A dialogue between the team and Donna Kennedy, Clinical Specialist in Hand Therapy at Imperial College Healthcare NHS Trust, was created and much of the success and direction of this project can be attributed to the expertise and insights that she offered. Donna believed there was scope for a device because rehabilitation currently involves repetitive tasks such as picking up beans and hundreds of therapist hours; the motivation is low and cost is high.

PNI refers to the damage of the nervous system outside of the brain and spinal cord. The particular focus of Michelangelo is for that of patients with upper limb damage, most commonly caused by trauma (such as motorbike accidents or lacerations caused by glass shards) (Kouyoumdijan, 2006). The areas innervated by the nerves lose motor control and sensory functions, thus inhibiting patients’ ability to perform activities of daily living; 2.8% of patients with PNI become permanently disabled (Rodríguez et al, 2004). The most severe classification of nerve injury is neurotmesis (Seddon,

1943), where all layers within the nerve are severed, requiring surgical repair. Following surgery, intensive therapy is required to regain hand function; the sooner the rehabilitation starts, the better the recovery, as sensory neuron death peaks at 2 weeks and motor neurons being to lose their regenerative properties after 6 weeks (Hart et al, 2008).

The main tenants of hand therapy are: • Support and protect the injured limb• Prevent joint stiffness• Maintain, facilitate and restore function• Desensitisation• Sensory re-education• Strengthening(Donna Kennedy, 2016a)

The primary focus of Michelangelo is the third tenant - restoring functionality in patient’s upper limbs, specifically the hand. The key focus of rehabilitation for patients is that of cortical retraining. One of the surprising insights from conversations with Donna was that movements are not indicative of recovery:

“Movement does not translate to movement. Just because someone can grip doesn’t mean there is any recovery. Strong fingers may be compensating for the weak ones. Muscles may not actually be firing. So detecting ‘a general movement’ may not be useful at all for a device. Sensory and cortical feedback is more key for quality of life; make the brain work by making it do specific task orientated problems to engage the brain” (Donna Kennedy, 2016b)

When a part of the body is immobilised, the corresponding cortical mapping in the somatosensory cortex shrinks (Lissek et al, 2009). Maintaining and expanding this mapping is the aim of therapy. The brain must relearn sensation and motor control to regain function, and contextual feedback provides cues to aid the process. In other words, our primary objective with Michelangelo is to provide individual finger feedback during repetitive tasks.

Further investigation demonstrated that patients with peripheral nerve injuries often develop Chronic Regional Pain Syndrome. The main symptom of CRPS is severe, continuous debilitating pain in the affected limb. Other symptoms include oedma (severe swelling) and skin discoloration, with hyperalgesia (increased pain response to stimuli) and allodynia (a painful response to an innocuous stimulus such as a cotton bud), and cold, hard surfaces producing more painful responses than soft warmer textures (Bruehl et al, 2002). Training weight bearing through the affected limb has been shown to reduce both swelling and pain, so a therapeutic programme of stress loading through the palm improves quality of life for affected patients (Carlson & Watson,

1988).

“Sensory and cortical feedback is more key for quality of life; make the brain work by making it do specific task orientated problems to engage

the brain”Donna Kennedy

Our primary focus: patients suffering from

PNI and CRPS

Page 1

concept generation2In order to narrow the focus and provide a beneficial solution that targeted a specific problem, rather than inadvertently create something broad-based but ineffective, the team began creating a hierarchy of priorities that would inform ideation.

Given the complexity of the hand, there are a plethora of motions that could be explored. Providing individual finger training and feedback was chosen as a primary focus. The movement of each finger can be well isolated and constrained. Feedback is more important than strength for PNI patients, so using simple buttons which require little force to press makes the device accessible to a wide spectrum of users, including more impaired patients who have just had surgery.

Another significant objective was making the design as inclusive as possible, to be of use to the widest spread of patients. This is evidenced by some our early concept sketches, as seen in Appendix A, and was successfully carried through to our final prototype.

In the same vein, creating a device that could be operated both independently and in conjunction with mobile computing devices was a strong objective. Whilst mobile technologies offer immense potential for tracking and engagement, our target demographic is older, so assuming ownership and familiarity with smartphones and tablets would be a mistake. A solution that could still provide benefits to all users without the need for a smartphone or tablet was very much sought after.

Further research and conversations with Donna brought us to the Dystrophile, a device designed to increase the weight bearing capacity through the patient’s palm (Carlson & Watson,

1988). Weight bearing through the palm is very important in daily living. For example, many elderly patients put their entire weight through their hands, pushing off of tables and chairs when trying to stand up (Kennedy, 2016b). The Dystrophile is a very expensive beige plastic brick to push down on, not ideal for those with cold pain.The current low-budget alternative is scrubbing tables with a towel for two minutes. We thought incorporating weight bearing through the palm in our product would address a gap in current devices and improve upon both the Dystrophile and the towel.

The objectives that dictated our prototyping can be summarised as follows:

• Provide a means for repetitive task orientated activities;

• Provide individual visual finger feedback;

• Quantitative measurement of stress-loading through the palm;

• Encouraging extension of the thumb to relax the hand and prevent hypercontraction;

• Create a device that can be used independently of mobile and computer technology;

• Ensure the design is inclusive and beneficial to the widest spectrum of patients;

• Create an ergonomic design that is aesthetic and motivates patients’ usage.

The dystrophile: expensive,

Plastic, not ideal for

those with cold pain.

Page 2

DESign analysis3The following images and annotations will provide a thorough breakdown of the various components in our final prototype. The path to our final model involved many iterations of prototyping, ranging from using putty to 3D printing parts and moulding silicone components. A full exploration and demonstration of the prototyping journey can be found in Appendix A.

ButtonsThese are easy to press, with an LED push button inbuilt into a silicone casing. The LED provides immediate visual feedback. The choice of silicone eliminates ‘cold pain’ for people with CRPS since it is both soft and warm, thus less painful for hypersensitive patients. The silicone covers can also be removed and washed, allowing for improved hygiene, particularly relevant if a singular device is being used in a clinical setting between many patients.

adjustable tracksThe buttons are on adjustable tracks such that they can be easily picked up and placed even by those with impaired hand function. This allows for the device to accommodate a range of different finger lengths and those whose palsies don’t allow for the full extension of their hands.

Body materialsThe device has a wooden top face which is soft and warm, minimising cold pain, provides strong aesthetic appeal, on top of a foam inner structure (which has been painted and sealed) and placed upon another acrylic face. All sharp edges have been removed and the device is lightweight so as to be easily transported by even the weakest of patients.

PALM PADThe palm pad locates the force sensitive resistor and allows for the interchange of different textured silicone covers to counter allodynia. An inbuilt LED gives a visual indication of the force through the palm pad, as it glows brighter as the patient presses harder.

Page 3

ElectronicsThe electronics were designed to match the layout of a human hand. Each finger button has its own breadboard that is connected to a central Arduino MICRO that lies underneath the palm of the device. Each finger is controlled in the following manner. The Arduino program first selects a particular finger using a random number generator. When the finger input to the Arduino is low (i.e the button isn’t being pressed) the Arduino output is set to high for that particular finger. When the button is pressed, the input is high, the LED is lit and the Arduino output is set to low. This mechanism is detailed in the circuit diagram shown in Appendix C. This game encourages users to individually focus on each finger as they ‘chase the light’. The LED provides clear visual feedback as to when the user is successful in pressing the button.

The palm circuit is a simple voltage divider circuit with a force sensitive resistor (FSR) and LED in series. The harder the user presses, the lower the FSR resistance, the greater the voltage drop across the LED, the brighter it shines. The voltage drop across the LED is measured by an Arduino analog port.

In all circuits the resistors between LEDs, switches and grounds are to protect the Arduino from current surges when the LEDs are lit. A sample of our Arduino code can be found in Appendix D.

APPWhile we have built the prototype to be independent of any mobile technology, we recognise the merits of using these devices and as such have created a simple interface by which patients can sync up the device with an app.

Blynk is an iOS application that can read, store and track data from a serial port connected to the Internet. Currently, to link the Arduino to the app we use a USB cable connected to a laptop. A bash script (vshymanskyy, 2016), combined with a multipurpose relay executed in the computer’s Terminal, connects the Arduino’s serial port to the computer’s WiFi network. Data from the Arduino is then transmitted to Blynk via the bash script.

The data transmitted to the app is private and can only be viewed by those with the appropriate authorisation code, keeping the patient’s sensitive health data secure. The app stores reaction times (an indication of finger isolation) and palm force and calculates daily, weekly and monthly averages to give an overview of progress.

Page 4

future improvements4The model presented at the end-of-project seminar merely marks the first substantial step for this device. There are considerable improvements and developments to be made before it can be a truly marketable product:

Electronic Improvements:• Adding a WiFi shield to the processor would allow Michelangelo to communicate wirelessly

over the Internet, removing the need for a USB cable (although this would likely remain for accessibility).

• Incorporating a lithium ion battery would allow Michelangelo to power itself, for greater independence and portability.

• Replacing the breadboards with printed circuit boards would make for sturdier and smaller connections, thus reducing the volume of the electronics and improving reliability.

• Changing the wires for conductive tracks, similar to those used in keyboards or sound mixing desks, to make the connections more robust.

Interface Improvements• The full realisation of a left and right handed operation. Michelangelo currently has a 6th

track to swap the thumb over, but is for demonstration purposes only.• The possible addition of extra components (an LED screen, speakers, vibration motors) to

provide greater sensory feedback and independence to the device.

Software Improvements• A greater range of applications or the device, fully utilising the range of sensors to create an

engaging experience. These may include games which are designed to complement the desired therapy program, although the scope is endless.

• A ‘front-end’ consumer facing suite of applications, games and history of progress which is separate from a ‘back-end’ user interface for therapists to update the device, reprogram it with patient specific tasks, view patient history, and set higher patient goals.

Mechanical Design Improvements:• Reduce the overall size and weight to optimise the device for use at home, by patients.• Improve the ergonomic casing to make it very easy to pick up and manoeuvre with one

working hand.• Redesign the casing and optimise the materials for mass-manufacture.

CONCLUSION5The primary motivation for this project was improving the rehabilitation of patients. By spending significant time researching literature and speaking to clinicians to really understand patients’ situations and the pains that they have to go through, we have created the seed for a potentially fantastic solution.

By bringing together our combined knowledge of mechanical, electrical and human-centred engineering design, our proof of concept prototype is a significant stepping stone to a fully realised, clinically beneficial device and has laid a foundation for any would-be commercial device.

The reception to the prototype has been very successful, with it winning the BE3-HHCARD Competition and praised by Donna Kennedy and her colleagues, who believe there is an excellent commercial potential for the device.

Page 5

REFERENCES6Bruehl, S., Harden, R. N., Galer, B. S., Saltz, S., Backonja, M. & Stanton-Hicks, M. (2002) Complex regional pain syndrome: are there distinct subtypes and sequential stages of the syndrome? Pain. 95 (1-2), 119-124.

Carlson, L. K. & Watson, H. K. (1988) Treatment of Reflex Sympathetic Dystrophy Using the Stress-Loading Program. Journal of Hand Therapy. 149-154.

Kennedy, D. L. (2016a) Traumatic Peripheral Nerve Injuries. [Lecture] BE3-HHCARD. Imperial College London. 22 Jan 2016.

Kennedy, D. L. (2016b) Interviewed by D. Gajjar-Reid, U. Hossain and S. Chaudhuri (4 Mar 2016).

Kouyoumdjian, J. A. (2006) Peripheral nerve injuries: a retrospective survey of 456 cases. Muscle & Nerve. 34 (6), 785-788.

Lissek, S., Wilimzig, C. , Stude, P. , Pleger, B., Kalisch, T. , Maier, C. , Peters, S. A. , Nicolas, V. , Tegenthoff, M. & Dinse, H. R. (2009) Immobilization impairs tactile perception and shrinks somatosensory cortical maps. Current Biology : CB. 19 (10), 837-842.

Rodríguez, F. J., Valero-Cabré, A. & Navarro, X. (2004) Regeneration and functional recovery following peripheral nerve injury. Drug Discovery Today: Disease Models. 1 (2), 177-185.

Seddon, H. J., Medawar, P. B. & Smith, H. (1943) Rate of regeneration of peripheral nerves in man. The Journal of Physiology. 102 (2), 191-215.

vshymanskyy. (2016) [Code] blynk-ser.bat GitHub repository. Available from: https://github.com/blynkkk/blynk-library/blob/master/scripts/blynk-ser.bat.

Page 6

AppendicES7APPENDIX A - Prototyping JourneyWe achieved a considerable amount of prototyping over the duration of the project.

Initial SketchesThe core idea was that of a palm pad and finger buttons laid out to fit under an outstretched hand. The following sketches show the various thought processes that brought us to buttons.

Page 7

Simple prototypingBy experimenting with mouldable materials, such as a glove filled with cornflour, homemade dough and plasticine, we gave shape and basic layout to the idea of a tactile interface. This was then replicated with more robust material: blue modelling foam, which in turn provided a mould for a resin prototype.

Initial Proof of Concept IdeaAt this stage, the scope of the device was fixed, with silicon buttons incorporating visual feedback instead of a flexible interface. Acrylic was laser cut and a basic wooden frame was created, to support and enclose the electronic components.

Page 8

CAD ModellingWith the idea crystallised through these early stage prototypes, the model was given more sophistication by replicating and refining it in SolidWorks. During this, factors such as inclusive design were considered and finalised, to ensure an adequate range of adjustability. These included the idea of having the buttons on movable tracks, and adding a 6th track that would incorporate another thumb button, thus allowing for left and right hand use. The final dimensions were also chosen.

Further PrototypingAs details of the design were refined, these were updated on the SolidWorks model to be eventually laser cut for the final prototype. The individual breadboards for each finger were made smaller and the housing was also simplified for ease of construction.

Page 9

APPENDIX B - Clincial RelevanceThe following is a table that summarises the features of the device and the clinical considerations that they are designed for.

Page 10

FEATURE CLINICAL RELEVANCE

Simple to press buttons Easy for all patients to press. Feedback is more important than finger strength

LEDs inside buttons Provides visual feedback to the brain

Left hand and right hand options Supports injuries on both sides

Washable buttons Hygienic for use between patients

Different programs Can isolate or exclude fingers based on requirements

Wooden face with no sharp edges Accommodates patients with CRPS to produce minimal pain

Force Sensitive Resistor in Palm Encourage weight bearing through joint

Variable palm pad texture Can increase difficulty and complexity for patients who improve

Quantitative progress indicators on app Provides motivation to beat own num-bers

Lifting/locking track Simple for those with functional im-pairments to adjust and move

Thumb extended Preferred clinical pose to prevent hy-percontraction of muscles

Auditory feedback Helps retain cortical mapping within brain through many feedback chan-nels.

APPENDIX C -Electrical Diagrams:The following is a set of diagrams to illustrate the circuits and code present in the final prototype.

FIGURE 1: CONTROL DIAGRAM FOR FINGERS

FIGURE 2: CIRCUIT DIAGRAM FOR FINGER PUSH BUTTON AND LED

FIGURE 3: CIRCUIT DIAGRAM FOR FSR AND LED IN PALM PAD

Page 11

APPENDIX D - ARDUINO SAMPLE CODEPSEUDOCODE:void loop() {• Begin serial connection • Connect Arduino to Blynk using authentication code• Run Blynk• Read analog sensor value from palm, convert this value to a voltage using the equation

voltage = sensorValue * (5.0 / 1023.0);• Pick a random number between 1 and 5, and set the corresponding finger LED Arduino

output to high. Start reaction time counter• While (corresponding finger input to arduino = LOW)LEDoutput=highEnd • end reaction time counter, calculate reaction time.• Write the palm voltage and the reaction time to Blynk virtual port. }

ACTUAL CODE#include <SoftwareSerial.h>SoftwareSerial SwSerial(2, 3); // RX, TX#define BLYNK_PRINT SwSerial#include <BlynkSimpleSerial.h>

// variables will change:

int ledPin=5;int buttonPin=4; int rando=0; int input =0; int start=0;int timend=0; int reaction_time=0; char auth[] = “de8cbf40af04455b899983bfa5815830”;int AO=0;

void setup(){ SwSerial.begin(9600); Blynk.begin(auth); // Default baud rate is 9600. You could specify it like this: //Blynk.begin(auth, 57600);}

void loop() { // Read the input on analog pin 0:

// Convert the analog reading (which goes from 0 - 1023) to a voltage (0 - 5V):

// Print out the value you read:

// Wait 100 milliseconds delay(100); Blynk.run(); int sensorValue = analogRead(A0);w float voltage = sensorValue * (5.0 / 1023.0); pinMode(ledPin, OUTPUT); pinMode(buttonPin, INPUT); digitalWrite(3,LOW); start=millis(); while (digitalRead(buttonPin) == LOW){ digitalWrite(ledPin,HIGH); //ensure pin[] and (func)button are different !!!!!! or else won’t work } timend=millis(); digitalWrite(ledPin,LOW);reaction_time= timend-start; rando= 2*random(1,6.00); input= (int) rando; buttonPin=input; ledPin=buttonPin+1; Blynk.virtualWrite(1, input); Blynk.virtualWrite(2, reaction_time); Blynk.virtualWrite(3, voltage); delay(500);

}

Page 12