herc neves - tu/e · herc neves principal scientist ... sor mclh lh so al mfb sm sm vlh hdb mfbb...

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).


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.


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.


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


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.


Technology timeline

20092007 2008 20102006


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


Modular integration

Original Michigan approach

Utah approach

NeuroProbes approach


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


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


Chemical sensing

• Intrinsic difficulty to integrate with in vivo probes• Constrained to two compounds of interest: choline and

glutamate


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.”


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


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


Depth control in NeuroProbes


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


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


Concept


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)


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


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


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


Control software

Control for data acquisition Data visualization Electrode selection

Plotttisettings

Status


Control software – electrode selection

Multifunctional microprobe arrays for the brain | 16th December 2010Slide 24 | www.neuroprobes.org 2

Control software – SNR visualisation


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)


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


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



passive probe

EDC probe

passive probe

sensory cortexmotor cortex



EDC probeEDC probepassive probe

sensory cortexsensory cortexmotor cortex


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


MUA

Cross correlograms of up-state onsetsat different channels

Cortex-cortex from the same shaft Cortex-cortex between shafts

40ms delay0ms delay


Cross correlograms of up-state onsetsat different channels

Thalamus-thalamus from the same shaft Cortex-thalamus between shafts

40ms shift0ms delay

Thalamus leads!


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


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


Acknowledgments

The European CommissionIstvan Ulbert (IP-HAS)Patrick Ruther (IMTEK)The NeuroProbes team

( )

For more information:


[email protected]

herc neves - tu/e · herc neves principal scientist ... sor mclh lh so al mfb sm sm vlh hdb mfbb...

Documents