Download Proposed SMD topology (cont`d)

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Opto-isolator wikipedia , lookup

Power engineering wikipedia , lookup

Power over Ethernet wikipedia , lookup

Switched-mode power supply wikipedia , lookup

Mains electricity wikipedia , lookup

Multidimensional empirical mode decomposition wikipedia , lookup

Alternating current wikipedia , lookup

Immunity-aware programming wikipedia , lookup

Transcript
Biomedical circuits and systems to recuperate
neuromuscular functions
OUTLINE
Biomedical circuits and systems to
recuperate neuromuscular functions
Main technological breakthrough
of FES based SMDs
w 1947, transistor allows circuit designs suitable for implants.
w 1952, first external pacemaker which was the size of a table
w Introduction
n
Mohamad Sawan, P. Eng., Ph.D., FCAE,
Purpose and motivation
radio of the time and was powered by 110 VAC.
w Microelectronics activities
Canadian Research Chair on Smart Medical Devices
n
PolySTIM Neurotechnology Laboratory
n
Dept. of Elec. Eng., École Polytechnique de Montréal
[email protected]
w 1957, electrical stimulation in the inner ear of the acoustic
Design, construction & test of smart medical devices (SMDs)
Typical architecture of an SMD
n
PART I
n
Neurostimulation to recuperate the bladder functions
Intracortical neurostimulation to create vision for the blind
Required biocircuits & biosystems
Research team and collaborations
Conclusions
IEEE CASS DLP
January 2004
2
M. Sawan
w
w
w
w
Pacemaker: More than 350,000 pacemakers annually;
Defibrillator: More than 50,000 implants annually;
Cochlear implant: More than 50,000 people worldwide;
Functional stimulation: Recuperate hands and legs
movements of paralyzed patients;
Stimulation for the treatment of Parkinson, respiration,
pains, scleroses, etc;
Automatic liberation of insuline in diabetics;
Neuronal activities recording for better control;
Devices for vision: Creation of adequate vision for blinds;
Other implantable microsystems.
Modulator
Power
supply
n
Low Pass Filter
Efficiency and safety.
Data transmission
n
n
Manchester based codec
PSK based codec: BPSK
Demodulator (COSTAS
Loop)
m(t)sin(w1t+q1)
• Recover the
carrier and the
data in the same
loop
-LSK: 200Kb/s
-BPSK: 1Mb/s
7
Without Feedback: 12.03%
With Feedback: 22.04%
controller
Power
Amplifier
On-Chip
C1
Charge Pump 1
Voltage
Regulator
L1
L2
Charge Pump 2
Vdd
Charge Pump 3
C3
C2
Charge Pump 4
Vdd1
Vdd2
Vdd3
Vdd4
Current sources
DEMUX
Signal
Processor
electrodes
Clock
Generator
Load Shift
Key (LSK)
Encoder
ADC
MUX
Analog
frontend n
Stimulation signals
Data
Demodulator
- ETC impedance measurement
- Charge per phase estimation
- ETC voltage monitoring
- Neural signal recording
5
Analog
frontend 1
Controller/
Stimulator
Electrodes
External Controller
Start-up
Circuit
Protection
Circuit
Clock Generator
Circuit
6
M. Sawan
• Lower urinary tract deficiency following
a spinal-cord injury;
• Conventional electrical neural stimulation (ENS)
induces contraction on both detrusor and sphincter,
thus preventing from complete voiding;
- Incontinence and/or retention.
• Conventional ENS combined with nerve rhizothomy is
not accepted by patients;
• Multiples complications at several
levels (Sphincter, bladder and kidneys)
due to repetitive catheterizations;
Q branch
• Transfer rate:
• PSK demodulator consumes only
529 mW.
•QPSK may double the data rate.
Off-Chip
Vdd
The bladder controller: Dual-stimulator
m(t)sin(q 1-q 2)
• Full duplex mode
Implant Circuit
Switching
Regulator
Data
Modulator
The bladder controller: Dual-stimulator
Data Out
m(t)sin2(q1-q 2)
2
Low Pass Filter
ASK Demodulator /
DAC/Decoder
Battery
Test
stimuli
Returned informations
Low Pass Filter
Phase
Shifter
90
data
Measurement
and
digitalization
circuitry
Data
processing
2cos(w1t+q2)
Link features
M. Sawan
VCO
Clk
Main
Physiological
stimuli
MUX
2sin(w1t+q 2)
Data In
Par
Stimuli generator
Back
telemetry
Demodulator
RF LINK to Transfer Power & Data
(cont’d)
h total = h rflink ⋅ h rectifier ⋅ h regulator
Receiver
Skin
Power Transfer
Global view h = h ⋅ h
total
rflink
rectifier ⋅ h regulator
Implant
External
controller
M. Sawan
I branch
m(t)cos(q1-q2)
Proposed SMD topology
(cont’d)
Bi-directional
RF link
Features:
4
M. Sawan
3
M. Sawan
Proposed topology
(Cont’d)
w
nerve in a totally deaf human was reported.
1960, totally implanted pacemaker (Buffalo).
1961, peroneal nerve stimulator for foot drop in hemiplegics.
1968, implantation of an electrodes array on the visual cortex.
1980, first microchip was used to design small pacemakers.
1984, FDA approved the first cochlear implant for adults
1991, recording of neural activities
1992+, FES is widely used in several applications.
Smart medical devices
Main technological breakthrough
w
w
w
w
w
w
w
w
w
w
w
Focus on two main case studies
• Early ENS tends to reduce the hyperreflexia phase,
while limiting damage in the whole urinary system;
• Bladder hyperreflexia (hypertrophy
or atrophy) hardly curable with
traditional medicine;
• Selective neural stimulation proposed by our
team allows efficient voiding of the bladder;
Electrical stimulation of the detrusor muscle allows partly
voiding and may induce kidney complication.
•
M. Sawan
8
• Permanent low frequency combined with
selective stimulation are proposed by our
team to improve bladder function.
M. Sawan
9
1
The bladder controller: Dual-stimulator
Selective neural stimulation
The bladder controller: Dual-stimulator
Selective neural stimulation
First selective stimulator includes :
l
• Selective stimulation allows improvement in voiding;
• This technique diplexes high and low frequency stimuli
to activate both categories of sacral nerve fibres:
The bladder controller: Dual-stimulator
Permanent/Selective stimulations
l
‡ Implant of 3.5 cm of diameter built around field programmable
Proposed dual-stimulator includes :
-Permanent LF stimulation : Low amplitude train:
- Prevent the bladder hyperreflexia
Lesion
- Maintain the bladder shape Skin
- Selective stimulation.
devices (FPD) and other discrete components;
‡ Shape Memory Alloy (SMA) based cuff electrodes to facilitate
connecting to sacral roots;
‡ User friendly external controller based on FPD and other
discrete components.
- HF stimuli for somatic fibres which innervate the sphincter.
- LF stimuli for parasympathetic fibres which innervate the
detrusor.
Selective
Implantable
Stimulator
User Interface
RF Emitter
l
Toff
- Battery to power the
permanent stimulator;
- RF link for the selective
stimulation.
1/Freq
Amp
11
M. Sawan
RF_PWR
Switch
#1
RF_PWR_EN
Antenna
Decoder
BAT_PWR
G_CLOCK
Controller
#1
(FPGA)
Bus
Controller
DIN/SCLK/CS
3
UP/DOWN
READY
DIN/SCLK/CS
Switch
#2
BAT_PWR_EN
2
3
Battery
SYS_ON
SYS_OFF
DSR
Switch
#3
Evolution
w 1960s, few groups began to investigate the possibilities of
exploiting the phenomenon of creating points of lights;
w 1990s, Researchers at NIH demonstrate that intra cortex
12%
Voiding by selective stimulation
UP/DOWN
2
SYS_ON
The visual cortical stimulator
Results
Intravesical and intraurethral pressures measurements and
Electromyographic (EMG) recording of the pelvic floor muscles were
performed.
l
SWT_PWR
Z_CTRL
Am
Demodulator
DATA_IN
MANCH_IN
12
M. Sawan
Bladder dual neurostimulation:
The bladder controller: Dual-stimulator
Implant description
250 ml
stimulation allows to generate phosphenes;
88%
PIC_PWR
Bladder
The system includes:
PW
10
Dual-stimulator
Implant
External Controller
Ton
M. Sawan
Bipolar cuff
Electrode
Power
Data
Param Stim #5
30Hz 175us
0.9mA
SMA Electrode Cuff
200 ml
Mean
Bladder 150 ml
Volume
PIC_PWR
w Progress in microelectronics and microfabrication motivate
Residual
researchers to explore several approaches;
66%
100 ml
Voided
34%
DSR
READY
SYS_ON
31%
BAT_PWR_EN
cath
SWT_PWR
3
UP/DOWN
2
DAC
Current
Source
Current
Commutator
n
n
14
M. Sawan
n
n
n
n
n
n
n
n
M. Sawan
n
n
n
n
n
n
n
n
M. Sawan
n
n
n
n
Fixed, stable phosphenes
Many phosphenes with few electrodes
Phosphene characteristics widely variable.
Tongue stimulation, and auditory or
tactile input.
17
Better resolution
Better control on phosphenes
Lower power
Lower chances of seizure.
Subject: 38 year old woman
Images feed from a digital camera to a beltmounted processor;
commands send through the skull, and into
the visual cortex
Optic nerve stimulation
Artificial retina requires functional retina and optic nerve;
Surface stimulation of the visual cortex is imprecise;
Other techniques only allow sensory substitution .
16
Intra-cortical stimulation
Stable, non flickering phosphenes;
Able to distinguish shapes;
Lower threshold in macular region;
Resolution restricted by electrode shape
Mechanical properties challenges.
Extracortical surface stimulation
n
Intracortical stimulation does not depend on the health of
the eye or optic nerve; also, it requires stimuli at low current
level. It can be the predominant one:
n
15
Intracortical stimulation
Experimentation
n
Artificial retina;
Cuff electrodes around the optic nerves;
Tongue stimulation;
Surface stimulation of the visual cortex;
Intracortical stimulation of the visual cortex.
Early 2000, a prototype has been completed to prove the
feasibility of a visual cortical stimulator
Miniaturized implant version is being achieved.
M. Sawan
Artificial retina / Extracortical
stimulation
Researchers worldwide are developing solutions
Several approaches are being explored:
n
cath
External controller
Subcutaneous Implant (~ 2.5cm)
One cuff electrode pair
Approaches
n
LF+HF
LF
+permanent
w We start this project in 1996
The visual cortical stimulator
n
LF
Nerve
13
M. Sawan
holding a PC which drives a percutaneous connector toward
the skull to extracortical visual region;
0 ml
SYS_OFF
DIN/SCLK/CS
•
•
•
w Dobelle institute, New York, early in 2000 presented a patient
69%
50 ml
RF_PWR_EN
Controller
#2
(PIC)
Parameters affect brightness, not position
Elicits stable phosphenes
Threshold ~15 µA (as low as 1.9 µA)
Short term accommodation.
Experiments in Monkey at IIT
n
M. Sawan
Recently, report good news.
18
2
The visual cortical stimulator
Integrated smart devices
Prototype (VCS)
Tx
Rx
Stimulator
• Block diagram
Interface
Timing Hardware
5
Antenna
6
Controller
Chip
n
n
19
20
M. Sawan
The visual cortical stimulator
n
n
Stimulator Modules (SMs)
Current Constant Stimulation
w Monopolar, Bipolar
w Single/Multiple paths
n
stimulator monitoring
w One to 25 bits per pulse.
Interface Module (IM)
Enables the use of a LF Clk
w System power saving
w Bidirectional wireless
Communication
n
n
w A/D Conversion of monitoring data
Interface Module
n
Analog
out
n
REF
n
Compatible with electrode
matrix dimensions
n
n
23
Downlink Communication
Downlink Tradeoff
VDD
High Q > Best efficiency, but
Sampling at any programmable time
Voltage measurement
n
Current monitoring
Coil Voltage
n
Digital decoder
I monit
SS
Current Monit.
REF
REF
j
Dig. Dir I n
Dig. Att In
w Toggling of attenuated carrier detected by
From
Ctrler
enable
adcclk
adccvt
To
Ctrler
adcout
Normal Stim.
+
-
Hi-Z
DAC
DAC
DAC
DAC
Voltage Monit.
Distributed
reference
Bipolar
Current
sink
Current
sink/source
Current
source
• Modular: Variable number of stimulation sites
R2R
• Stimulation modes: Mono / Bipolar, Single / distributed return terminal
REF
• Electrical Isolation: Allows stimulation current / voltage monitoring.
Hi-Z
M. Sawan
clk(N)
– Safety
ADC
extra toggling of direct carrier.
25
clk(1)
clk
– Flexibility
V DD
VSS
VthL
Lsk
clk
cvt
FSM
R2R
REF
j
VthH
V -V
w Based on carrier signal attenuation
M. Sawan
Distant
reference
Contr oll er
DD
ctrl / statusreg
errflg
Contr oll er
Analog
Monit. Bus
V +V
Direct Coil
Att. Coil
by reducing carrier OFF period.
data(N)
Monopolar
Hi-Z
than ~500 kbps @ 13.56 MHz.
data(1)
in/ out reg
LSK
out
Anodic/Cathodic
Stimulation
… Limited Data Rate
w Enhance Data Rate & Power Transfer
(toin/out FSMand
err c
trl / statusreg)
in/out
FSM
Stimulus mode/type
w Actual stimulation current measurement.
w Manchester encoding does not allow more
error
detec.
Cortical stimulation
w Any stimulation site during normal stimulation
limited by Q-factor of receiver when
rectifier diodes voltage is reversed.
n
n
n
w Envelop transition speed
data
clk
24
M. Sawan
V / I monitoring
ASKin
Manch.
decod.
Actual Stimulating Current
Voltage at every stimulation site
Monitoring & Data uplink
Power Efficiency vs Data Rate
n
n
M. Sawan
Clk
Clk
Data
Data
ext.
Env.
det.
Envelop env
detect.
Monitoring
w 4 x 4 electrodes (prototype)
PWR
data(N), clk(N)
Vmon(N)
err(N)
Controller
w Manchester LSK
n
RAM Configured Parameter
continuous error checking
LSKout
Uplink
Power management
In
Communication Error
detection/correction
Stimulator Instructions Dispatching
PWR
data(1), clk(1)
Vmon(1)
err(1)
LSK
mod
ASKin
System Control
Voltage Monitoring
w Disables stimulation in case of
22
F
RE
(Tx/Rx), but need higher data rate
w «All digital» Demodulator
n
Monit.
Regul
Downlink
w ASK: keep OOK for its simplicity
Band-Gap
Bias
R2R
External Coil, Rectifiers
Communication
Electrode
Switch
Matrix
Parameter corruption
M. Sawan
Implant
Electrodes
Stim.
Module
21
n
w Implant control
w Power recovery / management
Interface
Module
Implant architecture
DAC DAC DAC DAC
#0
#1
#2
#3
LS
Serial in
clk
Configurable Comm.
protocol
Display
Stim.
Module
Power supply
Features (flexibility/safety)
w Multi-channel stimulation
w Electrode-tissue and
n
Transmitter
Stimulator Module
Components of Multi-modules
Inverse
Stim. Site to
PVC Mapping
Display
Implantable
components
M. Sawan
Implant architecture
Implant architecture
PVC to
SSA
mapping
Ordering
Encoding
& timing
Low Level
Data Manag.
Front-end for efficient image processing;
Tx / Rx for power efficiency and data rate;
Power consumption, size, safety, biocompatibility;
Interface for shape and surface.
VDB source
Visual Field
Emulator
Artificial
Visuotopic
Map gen.
Image
fitting
SSA to
PVC
mapping
Low level
Pulse
sequencing
Implanted
Components
LS
M. Sawan
PVCS emulating
a 16x16 matrix
Matrix
of
electrode
s
&
101100010
1
Direct
Mapping
S kin
n
n
Config.
Interface
System level optimization:
Camera mounted onto a pair of glasses;
External controller to process the images and commands;
Bidirectional data & energy transmitted by RF link;
Implant to stimulate the region of the cortex controlling vision.
Programming GUI
Visuotopic
Map
...
n
Image
enhancement
...
n
Stimulator
Image
acquisition
...
n
Development
board
External
Components
The system includes 4 main components:
n
Isolation
Con tr oll er
3
Image Processor
& Power source
Compute
r
LS
1 Input
Image
Data collected from
experimentation
External components
Inputs
RF Data and Power
4
Wireless
Transmitter
...
Front-end
(Camera
& DSP)
F
MRE
7
Stimulator
Chips
adc
lck
adc
cvt
Cortex
2
Camera
The visual cortical stimulator
Undertaken research projects
The whole proposed system
adc
out
The visual cortical stimulator
26
M. Sawan
27
3
Stimulation Module (4
( x 4 Matrix)
•
Implant assembly
n
Modular
DACs
TEST
structures
CMOS 0.18 µm, ~60 000 Gates
BIAS
Downlink
n
> 1 Mbps @ 13.56 MHz, Duty Cycle = 67%
500 kbps @ 13.56 MHz, Duty Cycle > 85%
n
400 um
> 100 mW load capability; P (err) <
ELECTRODES
CONN / CTRL
n
MONITORING
R2R AMP
Amplitude, Pulse width,
Inter-phase delay,
Direction of first pulse,
Pulse frequency, Pulse
train duration, and Intertrain delay.
Physical Dimensions
Wireless
Data rate >1 Mb/s
Sufficient for 1000 stimulation sites
n
Parameters
IM
Power/data link
CTRL
Power: <1mW/SM @ 1MHz
n
One controller &
several stimulation
Units
Collaboration with the INM
SMs
Flexible substrate
Uplink : 200 kb/s
10 -6
In vivo testing
Cortical stimulation
Implementation Results
256 stimulation patterns @ 50 Hz.
n Max. Stim. Freq. :
> 1 kHz @ 500 Elec.
I / V Monitoring
n
SM’s area : 3.5x3.5 mm2
n
Substrate : 50 µm
n
Stim Assembly: 1 mm thick
n
Electrode Length: 2 mm
n
IM’s Area : ~ 2 x 2 cm2
n
Ctrler thickness : ~ 3 mm.
Comput
28
Implant
M. Sawan
Downlink
Timing
Electrode matrix
400 µm spacin;
Platinum tips; ParyleneC coating, Low
Impedance.
n
Physical Properties
Safety
Data rate
Stimulation Flexibility
Power efficiency
n
Placement flexibility, wiring
Chronic usage
High number of stimulation sites
Prototype
High parallelism
n
n
n
Acquired Image
n
n
Irregular
- Density
- Size
- Sensitivity
-…
n
Fit
34
weighted with Schwartz retinotopic relationship
and estimated electrode placement.
n
Zones
Probabilistic weighted PVFC
variations and phosphene
parameters:
33
M. Sawan
Image acquisition/processing
Power Optimization
Fovea
Required stimulation
current for 1 sample frame
Min. Current
Summary
n
n
Fixed
Scan average
35
Design of dedicated image sensor
High voltage stimulation techniques
Modeling, simulation and in vivo validation
Measurement techniques of neuronal activities
Assembly several arrays
n
128, 500, and 1000+ microelectrodes.
Nanoelectronics will allow the design of additional devices.
Microelectrodes & power sources are among the major works.
Such medical devices include validation and involve physicians,
engineers, health care professionals, families, etc.
Max. Current
M. Sawan
Stimulation in the visual cortex may soon permit true artificial vision
Other undertaken research topics:
w
w
w
w
w
Total stimulation current
Adaptive
Scan average
Edges
Typical topology of inductively coupled implantable electronic devices
Dual-stimulator to recuperate the bladder functions
Main issues of a visual cortical stimulator
With large electrode arrays, stimulation current represents a major part
of power consumption
n Predetermined image scanning results in large current variations
n Adaptive scanning method regulates current demand
Fitting example
M. Sawan
Original
w Set of Phosphene Visual Field Coordinates
w Generated from probabilistic parameters
w Equalize the required power and reduce the peak demands
Horizontal Angle
or from
Realistic Visuotopic Map
w Sensitivity, Size, Brightness, etc.
Vertical Angle
n
w Investigation of representation needed for a VCS.
n
32
M. Sawan
Sample Visuotopic Map
Image has to be enhanced for
clarity, then mapped to the
available PVFCs.
The Visual Field Emulator allows
perception prediction / evaluation
from the research team.
Experimental determination
Creation of random maps :
Constructed from experiments
Visual Data Base Generator
Minimize power consumption with other techniques than circuit
optimization approaches.
Visual Data Base (cont’
(cont’d)
n
n
Present image as the patient would see it;
Hardware implementation of processing algorithms;
Low power consumption topologies.
Image acquisition/processing
Enhancement/mapping
n
Actual Perceived Image
(Direct Fit)
31
M. Sawan
Allow to transform images over visuotopic maps;
Follow a bottom-up approach until percepts are understandable;
Must deal with every image encountered in real world;
Be safe for patient.
Required processing
Low resolution
Control
on phosphene
parameters
Electrically invoked phosphenes
w Visuotopic Maps
Dedicated image processor
External Control
w Relevance of Data
Image acquisition/processing
Main specifications
Challenges and Need for Optimization
n Stimulator
Subject
(Monkey)
30
M. Sawan
Image acquisition/processing
Image acquisition/processing
w
w
w
w
w
29
Stimulator
Only electrodes in
subject; Configurable
hardware; “Large” output
voltage swing.
Stim. Units.
M. Sawan
Oscilloscope
Development
board
Assembly
n
PWR Data
Isolation
er
t
M. Sawan
36
4