herc neves - tu/e · herc neves principal scientist ... sor mclh lh so al mfb sm sm vlh hdb mfbb...
TRANSCRIPT
Multifunctional microprobe arrays for the brainfor the brain
TUEindoven, 16th December 2010
Herc NevesPrincipal Scientist Biomedical Microsystems imec (Belgium)Principal Scientist, Biomedical Microsystems, imec (Belgium)Professor of Physics, Uppsala University (Sweden)Coordinator of NeuroProbes
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 0 | www.neuroprobes.org
Neural recording
It is the only possibility for investigating the link between neuralensemble activity and subject behaviourensemble activity and subject behaviour.
In vivo studies aim to observe single-unit activity by measuringextracellular potentials.
Extracellular recordings collect signals from all the nearby neurons Extracellular recordings collect signals from all the nearby neurons(field potential). This usually has to be resolved into single units(action potentials).
For this reason electrodes have to be sufficiently close to neurons For this reason, electrodes have to be sufficiently close to neurons(typically less than 100 μm).
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 1 | www.neuroprobes.org
In-vivo cerebral recordings in humans
Modern non-invasive techniques such as fMRI cannot provideinformation on physiological processes in the brain at the cellular level.Clinical use of brain recording mainly in two areas:
Management of movement disorders resulting from pathologiesaffecting deep brain structures (thalamus, basal ganglia);
Localisation of pathological activity foci for surgical excision inintractable epilepsy.
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 2 | www.neuroprobes.org
The NeuroProbes project
The problem:p To electrically record three-dimensional neuron ensembles in vivo
over long implantation periods, also in monkeys and humans; To integrate such recordings with chemical sensing and drug To integrate such recordings with chemical sensing and drug
delivery; To perform electrical stimulation; To provide telemetry where needed To provide telemetry where needed.
The solution: A new 3D technology that permits a combination of diverse
functionalities; Embedded, distributed electronics;, ; Extra thin backbone, conducive to floating operation.
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 3 | www.neuroprobes.org
The NeuroProbes IP
ObjectiveDevelopment of arrays of multifunctional microprobes for high temporal and spatial resolution brain studiesthat include freely moving subjects, with recording and stimulation done both electrically and chemically.that include freely moving subjects, with recording and stimulation done both electrically and chemically.Features• Dense 3-D microelectrode arrays for recording and stimulation• Modular technology• Microfluidics for inactivation studies• Microfluidics for inactivation studies• Individual depth control for accurate placement• Attachment and insertion technologies• Conformation to convoluted surfaces such as sulci of highly folded cortices
I t t d bi b• Integrated biosensor probes• Telemetry
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 4 | www.neuroprobes.org
Key differentiators
Modular approach. Probes are assembled in a Lego® fashion that permits tobi b i h diff f i li i h d icombine probes with different functionalities, shapes and sizes on a
common backbone.
Chronic use. A combination of diverse strategies (not limited tobiocompatible coatings) for the management of foreign body reaction wasinvestigated in NeuroProbes.
Electronic depth control. Instead of relying on trial and error, our probes aredesigned to allow the individual control of electrode position with respect todesigned to allow the individual control of electrode position with respect tosingle neurons.
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 5 | www.neuroprobes.org
Technology timeline
20092007 2008 20102006
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 6 | www.neuroprobes.org
Boundary conditions
Recording in rats and monkeys Focus on the recording of single action potentials Possibility for chronic use (floating operation) Dual character of the project (technology & science): devices
had to be available to users very early into the project Relatively short duration of the project: limited set of Relatively short duration of the project: limited set of
variations, parallel paths, additional concepts Flexibility to integrate diverse functions into the same platformy g p
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 7 | www.neuroprobes.org
Modular integration
Original Michigan approach
Utah approach
NeuroProbes approach
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 8 | www.neuroprobes.org
Platform concept
Key point: electric and fluidic interconnect between off-plane probe y p p pand in-plane platform
Focus: reliable electric interconnect, robust fabrication process
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 9 | www.neuroprobes.org
Drug delivery
• Of particular interest in inactivation studies• A number of significant challenges: very small volumes, potentially
large dead volumes, difficulty to control dosage, need for aminiaturized pumping scheme, rechargeabilityp p g , g y
PMP-NC² NeuroMedicator Fluidic microprobes
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 10 | www.neuroprobes.org
Chemical sensing
• Intrinsic difficulty to integrate with in vivo probes• Constrained to two compounds of interest: choline and
glutamate
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 11 | www.neuroprobes.org
Challenges of in vivo extracellular recording
According to J.C. Lilly, in “Correlations between neurophysiological activity inh d h b h i i k ” Bi h i l B f B h ithe cortex and short-term behavior in monkey”, Biochemical Bases of Behavior(1940),
“… one of the large difficulties in correlating structure, behavior, and CNSactivity is the spatial problem of getting enough electrodes, and small enoughelectrodes, in there with minimal injury. Still another difficulty is the temporalproblem of getting enough samples from each electrode per unit of time over aproblem of getting enough samples from each electrode per unit of time, over along enough time, to begin to see what goes on during conditioning or learning,especially when a monkey can learn with one exposure to a situation, as we seerepeatedly. As for the problem of the investigator’s absorbing the data – if he hasadequate recording techniques, he has a lot of time to work on a very shortrecorded part of a given monkey’s life.”
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 12 | www.neuroprobes.org
Need for depth control
• All multielectrode probes rely on trial and error for positioning. There isno way to optimize electrode positioning of multiple electrodes in the sameno way to optimize electrode positioning of multiple electrodes in the sameshaft
• Need to adjust electrode position with respect to cells• Need for fine adjustments in the course of an experiment• Need to reconfigure experiments
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 13 | www.neuroprobes.org
Depth control approaches
I. Szabó, J. Neurosci. Met. 105 (2001) 105-
J. Muthuswamy et al., IEEE Trans. Biomed. Eng., 52 (2005) 1748-1755
, ( )110 J.G. Cham, J. Neurophysiol. 93 (2005) 570-
579
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 14 | www.neuroprobes.org
Depth control in NeuroProbes
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 15 | www.neuroprobes.org
Mechanical actuation – factors against it (at least in our case)
Mechanical displacement moves all electrodes in tandem
Needs a point of support (not available in floating operation)
Not good if tissue reaction is to be minimized
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 16 | www.neuroprobes.org
Concept evolution
3D integration technology
Multi function integrationMulti-function integration
Mechanical depth control
Conformation to complex surfaces
Electronic depth control
L b f ki l t d• Large number of working electrodes
• Extra control and heavier computational power needed
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 17 | www.neuroprobes.org
Concept
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 18 | www.neuroprobes.org
Basic architecture
CMOS wafershaped by post processing
Interconnect matrix:
shaped by post-processing
FF FF
FF FF
Contact fingers for platform or bond pads for flex cable
Large shift register (daisy chain)
switch on-resistance << electrode impedance
for flex cable (defined by post-processing mask) Electrodes:
- arranged in 2 columns to 8 parallel analog output lines per shaft can measure max
allow tetrode configurations- electrode size, shape and column pitch determined by
t i k
per shaft can measure max. 8 electrodes at same time per shaft (but electrodes can also be connected in parallel)
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 19 | www.neuroprobes.org
post-processing maskbe co ected pa a e )
Front-end amplifier
• 37 dB Harrison amplifier for each output line • 5mW overall power consumption37 dB Harrison amplifier for each output line (0.1Hz – 8.5kHz bandwidth)
• 200kHz 8:1 MUX (25 ksamples/s/channel)• Class AB output buffer (526kHz BW, 1.4V/µs SR)
5 o e a po e co su pt o• 3.7µVrms total input referred noise
(1Hz – 10kHz bandwidth): lower than electrode noise
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 20 | www.neuroprobes.org
Interface box
Digital I/O:Interface electronics (IMEC)CLK_INData IN
D t OUTDATA_IN
Data OUTClock
Electronic depth control signals
E t l l t d
Controller
Reference & biasDATA_OUT
External refSextSint
4x(1/row)
External electrode connections:System ground
Ground
External ref“Single ended”
Reference & biascontrol
…
128 analog outputs
DEMUX+ sample&hold
2nd
stage gain
16 Recon-struction
filter
…
ed utpu
tsks
ps
MUX clock 256kHz = 8x32kHz
128 analog outputs(possibly 16 directanalog outputs)
…
utpu
ts4x4 array = 16 needles8 channels/needle16
MU
X’e
anal
og o
u16
0-20
0
16 d
irect
an
alog
o
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 21 | www.neuroprobes.org
Implementation
R f i14-pin headerto interface boxReference pin
Ground pinto interface box
8mm long comb wire bonded to PCB Connector for direct
electrode access
CMOS front end
MEMS motherboard
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 22 | www.neuroprobes.org
Control software
Control for data acquisition Data visualization Electrode selection
Plotttisettings
Status
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 23 | www.neuroprobes.org
Control software – electrode selection
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 24 | www.neuroprobes.org 2
Control software – SNR visualisation
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 25 | www.neuroprobes.org 2
Slow wave activity in the thalamocorticalsystem
V. Crunelli and S.W. Hughes, “The slow (<1 Hz) rhythm of non-REM sleep: a dialogue between three cardinal oscillators” Nature Neurosci 13 9 17 (2010)
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 26 | www.neuroprobes.org 2
oscillators , Nature Neurosci., 13, 9-17 (2010).
Target identification
5 4 3 2 1 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 -13 -14 -15
Bregma
8mm EDCprobe
8
9
10
2
1
0 Bregma
RSA
V1MV2MM
fmj
CA1
PtAS1Tr
S1HL
M1
sensorycortex
6
7
4
3
Crus2apmf
icf
Crus1
6
6
psfSimprf
5
4
pcuf
DCIC
CIC
DpG
InWh
InGOpSuGZo
RSA
Shf
CA1
CA2
CA3 Hil
DG bsc
OTPPT
OPTLPMRLDfi
LV
cc
cg alv PostRSGa
APTDfmi
lo
rf M2
AOB
FrA
LO
3
4
5
7
6
5
Cop
ppf
PM3
2
pcuf
IntA
IntPscp
LPB
ECIC
CnF
InCo
pGLDAVDM
AVVL Ang
APTV
Eth Sc
SPFPC
Po
VPMPaR
DpMe4n
vsc
LR4VSuVe asDC
LVe
PCTg
RtVL
VA
ic
ic
stBSTMPM
BSTLAcbCAOV
lo
AOLaci
C1
CPuPir
AOD BSTS
MiTg
DEn
VO Thalamus(nRT)
1
2
3
9
8
7 MPBVM F
ZIml RRF
PaR
PPTg
SuVe as
X ECuCu
SpVe
PCRtA PCRt
Me5me5
Su5
Mo5
7n
ocb
sp5
Ge5DMSp5
Sp5CMdD
rsERS
VLTgtfp
lfpcp
SNR
SNCD
PBP
SThal
SubCV
PnO
SubCD
A7P5
SubI
mfb
MTuSOR
MCLHLH
SO
al
mfb
sm
SM
BSTL
VLHHDBmfbb
LPO
SIB
VP
VP
IPAC acpaca
AcbC
AcbSh
CBmfba
ICjTu
PirAOP MVeMC
a
SolVLSolIM
IRt
-1
0
11
10
Lateral 1.90 mm
Interaural
InterauralFigure 83
LRt7
P7LPGi
Amb
LSO
RPO
MVPO
tz
VLLll
PnAPit
tfpSORsoxoptSO
3n
a
RVLCVL
C1
IRtDPOA5
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 27 | www.neuroprobes.org 2
14 13 12 11 10 9 8 7 6 5 4 3 2 1 -1 -2 -3 -4 -5 -6 0
Experimental approach
EDC probe
sensory cortexmotor cortex
sensory cortex
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 28 | www.neuroprobes.org
Experimental approach
passive probe
EDC probe
passive probe
sensory cortexmotor cortex
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 29 | www.neuroprobes.org
Experimental approach
EDC probeEDC probepassive probe
sensory cortexsensory cortexmotor cortex
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 30 | www.neuroprobes.org
Acute recordings, 2D probes
sive
pro
beto
r cor
tex
sive
pro
beot
or c
orte
x
Pas
mo
robe
cort
ex
Pas
mo
prob
em
us
LFP
EDC
pse
nsor
y
EDC
pth
alam
e pr
obe
cort
ex
e pr
obe
cort
ex
Pass
ive
mot
or c
e tex
Pass
ive
mot
or c
e
MUA
EDC
pro
bese
nsor
y co
rt
EDC
pro
beth
alam
us
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 31 | www.neuroprobes.org
MUA
Cross correlograms of up-state onsetsat different channels
Cortex-cortex from the same shaft Cortex-cortex between shafts
40ms delay0ms delay
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 33 | www.neuroprobes.org
Cross correlograms of up-state onsetsat different channels
Thalamus-thalamus from the same shaft Cortex-thalamus between shafts
40ms shift0ms delay
Thalamus leads!
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 34 | www.neuroprobes.org
Work in progress
Analysis of the electrophysiological data (e.g. cross correlation peak )maps)
Histological reconstruction Histological reconstruction
Correlation of electrophysiology data with the anatomy
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 35 | www.neuroprobes.org
Some conclusions
Quite a bit of work for a 4 ½ year project
Short duration inevitably led to technology compromises
Significant achievements in system integration and development of electronic depth control
Many ideas for follow up
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 36 | www.neuroprobes.org
Acknowledgments
The European CommissionIstvan Ulbert (IP-HAS)Patrick Ruther (IMTEK)The NeuroProbes team
( )
For more information:
Multifunctional microprobe arrays for the brain | 16th December 2010Slide 37 | www.neuroprobes.org