Prototype fMRI compatible actuators Final version€¦ · Moreover we tested the final prototype of the haptic interface to analyze its compatibility with the MR environment and we

Prototype fMRI compatible actuators

Authors D. Chapius, R. Gassert, V. Hartwig, N. Vanello, A. Bicchi

Date 30th May 2004

Del./Task Identifier D4.11/T4.5

Work Package WP4: New generation of force feedback devices

Partner(s) UNIPI

Work Package Leader UNIPI

Confidentiality Level Public

Abstract:

The purpose of this document is to describe an fMRI (functional Magnetic Resonance Imaging)

compatible haptic interface to investigate the mechanisms of tactile perception within an MR

environment. This interface allows one translation with force-feedback along a horizontal axis as

well as one rotation about a vertical axis linked to the translation. It can be used to move and orient

various objects below the finger. The interface and its components MR compatibility properties are

analyzed in this report.

TOUCH-HapSys Prototype fMRI compatible actuators

2

TABLE OF CONTENTS

PROTOTYPE FMRI COMPATIBLE ACTUATORS 1

1 INTRODUCTION 3

2 2DOF MRI/FMRI COMPAT IBLE HAPTIC INTERFACE WITH PASSIVE AND ACTIVE

ACTUATION 4

2.1 Concept 4

2.2 MR safety and compatibility 4

2.3 Passive and active sensing and actuation 4

2.4 Working principle of the 2DOF MR compatible haptic interface 5

2.4.1 Actuation/Transmission 5 2.4.2 Sensing 6 2.4.3 Control 6 2.4.4 Safety 6

3 COMPATIBILITY TEST AND RESULTS 8

3.1 Preliminary compatibility test 8

3.1.1 Results 10

3.2 Haptic interface compatibility test 16

3.2.1 Results 16

3.3 Discussion 19

4 CONCLUSIONS 21

5 REFERENCES 22


3

1 Introduction Functional brain exploration methodologies, such as functional Magnetic Resonance Imaging

(fMRI), are at present used to study perceptual and cognitive processes. To develop more complex

experimental fMRI paradigms, researchers are interested in realizing active interfaces, using

electrically powered actuators and sensors to be used inside the MRI environment. The use of non-

ferromagnetic metals with higher stiffness and rigidity compared to plastic facilitates the design of

smaller devices. Several reports provide criteria for MR compatible devices[1][2].

Robotic systems working within an MR environment require the development of MR safe and

compatible actuators and sensors. There must be no mutual interference between the materials,

sensors and actuators and the scanner. Only non-ferromagnetic materials can be used, as these

would pose a severe safety threat within the strong static magnetic field of the scanner. Conductive

materials may be used if placed at suitable positions [3].

In this work we describe an fMRI compatible haptic interface to investigate the brain mechanisms

of tactile perception using active as well as passive MR compatible actuators and sensors: the

compatibility of these devices was studied in the D. 4.10 and here is further analyzed using a new

statistical test.

The compatible haptic interface was developed to investigate the mechanisms of tactile perception

within an MR environment. This interface allows one translation with force-feedback along a

horizontal axis as well as one rotation about a vertical axis linked to the translation. It can be used to

move and orient various objects below the finger.

This report describes the preliminary compatibility tests carried out at UNIPI, the MRI/fMRI

compatible haptic device that was developed based on these tests and discusses the MR safety and

compatibility of this device. Moreover we tested the final prototype of the haptic interface to

analyze its compatibility with the MR environment and we reported all results in this paper.

The MRI/fMRI compatible haptic interface was developed by EPFL as subcontractor of ETH.


4

2 2DOF MRI/fMRI compatible haptic interface with passive and active actuation

2.1 Concept The MRI/fMRI compatible interface developed in this project is based on the MR compatible robot

technology introduced in [4]. This technology assures a maximum level of MR safety and

compatibility, which is critical as this device will be used by a human subject while inside a

magnetic resonance imaging (MRI) scanner.

2.2 MR safety and compatibility Robotic systems working within an MR environment require the development of MR safe and

compatible actuators and sensors. There must be no mutual interference between the materials,

sensors and actuators and the scanner. Only non-ferromagnetic materials can be used, as these

would pose a severe safety threat within the strong static magnetic field of the scanner. Conductive

materials may be used if placed at suitable positions.

2.3 Passive and active sensing and actuation While the technology presented in [4] uses solely passive actuators and sensors inside the scanner

room (a master-slave system with a commercial DC torque motor placed outside the scanner room

actuating an MR compatible slave over a hydrostatic transmission, and sensors based on intensity

measurement of reflected light), the new interface presented here also uses active components.

Active components require electric energy to be sent into the MR room to generate mechanical

motion or power a sensor that will produce an electric output signal. We have investigated the

following additions with respect to [4]:

• Non-ferromagnetic metals for the slave piston. This helps reducing friction in the

hydrostatic transmission.

• An ultrasonic motor to power the rotary degree of freedom of the haptic device.

• A linear potentiometer to measure the position of the slave piston.

All of these additions have one thing in common: they require compatibility testing; as the metals

are electrical conductors and the ultrasonic motor and linear potentiometer require electric and

electronic signals from the control room into the MR room and back. In the case of the non-

ferromagnetic metals, eddy currents will be induced during motion in the fringe field of the scanner

or by the switching magnetic field gradients within the scanner bore. This leads to thermal and


5

mechanical effects and can disturb the imaging. The electric cabling required for the ultrasonic

motor and the linear potentiometer acts as an antenna and can pick up the radio frequency (RF)

signals emitted by the MR scanner or disturb the imaging. Therefore, these components can only be

used in specific locations (e.g. at a minimum distance from the scanner bore), and require MR

compatibility testing. The tests carried out to verify the compatibility of these components are

described in section 3.

2.4 Working principle of the 2DOF MR compatible haptic interface

2.4.1 Actuation/Transmission For this interface, a master-slave system with hydrostatic transmission is used to actuate the linear

degree of freedom. This concept was introduced in [4] and its performance was analyzed in [5]. The

original transmission was miniaturized and now uses single ended pistons made from brass. The

slave cylinder can be disconnected from the master cylinder to allow easy installation of the

interface within the MR facility. The transmission length is eight meters. The master hydraulic

cylinder is actuated by a DC torque motor over a differential belt and pulley system (Figure 1). The

rotary degree of freedom is actuated by an ultrasonic motor (USR60-E3N, Shinsei Corp., Japan) [6].

The slave system in integrated into a glass fiber box (Figure 2).

Figure 1 The master actuator: DC torque motor (a), master hydraulic cylinder (b), hydraulic pump (c) and circuitry (d) as well as disconnectable hydrostatic transmission (e) and security switches (f)

a b

e

c

d

f


6

Figure 2 The slave module: A) open version showing the slave hydraulic piston (a) and the tactile pad (b). The ultrasonic motor is located beneath the tactile pad. B) The closed module showing the touch-pad (c) and the

disconnectable (d) hydrostatic transmission

2.4.2 Sensing The master DC torque motor is equipped with a commercial quadrature encoder with 5000

increments per revolution. The displacement of the slave piston is measured using a linear

potentiometer mechanically linked to the slave piston. The ultrasonic motor is also equipped with a

quadrature encoder.

2.4.3 Control The interface is controlled by a commercial PC running Windows and developed control software

created with LabWindows from National Instruments. The hardware is controlled over a

Multifunction DAQ card from National Instruments connected to a custom designed

interface/security PCB. This PCB holds the interfacing electronics that convert the encoder signals

to the requirements of the acquisition card as well as the logic circuits that monitor the security

hardware. Power sources, safety hardware and the interface/security PCB are contained in a master

rack (Figure 3 A). Control is done at 500 Hz, which is sufficient for control of the used transmission

[5].

2.4.4 Safety As this interface will be used by a human subject within an MR environment, safety is a crucial

factor. The developed interface presents the safety features proposed in [4], including:

A

b

a

B

c

d


7

• A master emergency button (which can be disconnected from the master rack) for the

experiment supervisor to cut power to the DC and ultrasonic motors

• A security bellow for the subject in the MR scanner to disable the DC and ultrasonic motors

and avert the experiment supervisor over a pneumatic hose and switch (Figure 3 B)

• Electrical end-of-travel switches on the master actuator to disable the DC torque motor if the

master piston moves past the desired position (these positions can be adjusted by moving the

switches)

• Mechanical end-of-travel limitations (intrinsic property of the hydraulic cylinder)

• Software security routines which monitor position and speed of the DC and ultrasonic

motors

• A master enable button located on the master rack to disable the DC and ultrasonic motor

If any of these security features are activated, the green enable button on the master rack is

deactivated and a red security lamp lights up to avert the experiment supervisor.

Figure 3 A) Master control rack with emergency button and B) MR compatible safety bellow

A

B


8

3 Compatibility test and results

3.1 Preliminary compatibility test

Tests were carried out at UNIPI, in order to evaluate the compatibility of the materials and the

devices that can be used to realize the haptic interface and to provide criteria for its design: the

influence of various metal pipes (non-ferromagnetic metals as aluminum, copper and brass,) placed

just outside and inside the scanner bore (moving in translation as well as resting) were tested.

Moreover, we tested a shielded cable that is intended to be used with a linear potentiometer and two

types of motor: a DC motor and an ultrasonic motor.

Likewise, the ultrasonic motor was placed at the scanner bore entry and tested at rest and while it

was actuating a mechanical brake, i.e. while it was loaded and producing high power.

Across all experiments, we scanned (by means a scanner Signa Horizon 1.5T, GE Medical Systems)

a spherical phantom of CuSO4 solution, using a GE-EPI (gradient echo, echo planar imaging) with

the following parameters: TE/TR 40/3000 msec, bandwidth 62.5 kHz, FOV 24 cm, resolution

64x64 pixels, Flip angle 90°, Slice thickness 5 mm, number of slices 25, 25 time frames long

sequences acquired.

The Signal to Noise Ratio (SNR):

SDcornerPcentreSNR /= (1)

where Pcentre is the mean value of a 11x11 pixels area at the centre of the image, and SDcorner is

the standard deviation of a 10x10 pixels area at the higher right corner [7], and the standard

deviation (SD) of each voxel signal in time domain were calculated.

We used a t-test in order to look for differences between the above parameters regarding image sets

acquired in various experimental conditions and image sets acquired with no device (reference

images)[8] .

The Student t-test is a parametric one (the hypothesis are based on the distribution parameters, i.e.

mean and variance) that assesses whether the means of two groups are statistically different from

each other.

The hypothesis can be summarized this way:

-null hypothesis 210 µµ −=H

-monodirectional alternative hypothesis 211 µµ >=H

-bidirectional alternative hypothesis 211 µµ ≠=H


9

where 1µ and 2µ are the means of the two populations.

If the number of elements 1n and 2n of the two samples are smaller than 30 (in our case the

number of slices for each acquisition sequence is 20) the equation for the t-test, under the a priori

hypothesis of equal means, is the following:

+−+

+

−=

21

21

21

222

211

21

2

)(

nnnn

nnsnsn

xxt (2)

where 1x and 2x are the means of the two samples and 21s and 2

2s their variances.

In the previous equation, the upper part of the ratio is just the difference between the two means or

averages. The lower part is a measure of the variability or dispersion of the scores.

This indicator has 221 −+ nn degrees of freedom.

Equation (x) is valid when the two statistical populations have the same variance. This condition

may be not fulfilled in our case, thus the denominator of the equation must be replaced with:

11 2

22

1

21

−+

− ns

ns

(3)

The t-value will be positive if the first mean is larger than the second and negative if it is smaller.

In order to decide whether the null hypothesis is true or not we must fix a critical value for the

indicator t: it is possible to associate the critical value for the probability of the null hypothesis to be

true, given the number of degrees of freedom and the kind of alternative hypothesis (in our case the

alternative hypothesis is bidirectional ), with the critical value for t.

If the estimated t-value is larger than the critical one then you can conclude that the difference

between the means for the two groups is different.

If we choose that the significance level equals to 0.05 (this means that five times out of a hundred

you would find a statistically significant difference between the means even if there was none) and

considering that in our case the number of degrees of freedom is 38, we get a t critical value of

2.024.

If we apply the t test in order to evaluate significant differences between acquired sequences with

the devices under test and reference images and we find a t value greater than 2.024, we can assert

that the relative experiment created significant artefact in the images.


10

We hypothesized that the parameters’ variance in each set could be different due to the motion of

the objects, or to the motors or currents being turned on and off to simulate the devices’ working

conditions.

The t test has supplied results more stable and coherent than ones used previously and this allowed

to obtain information also about the stability of the measurement system during the same session of

tests and sessions made in several days: we acquired in fact several reference sequences (same

investigation conditions) and compared them using the statistical test so as to highlight possible

differences owed to the system itself. Distinguishing the artefact owed to the devices under test by

the ones due to instability of the scanner in a less ambiguous way is possible so.

3.1.1 Results The following tables show the results obtained by the statistical test used for last analysis (t test) for

the three metal pipes, the ultrasonic motor, the DC motor, the cable and the linear potentiometer,

and finally, the scanner temporal stability.

Statistically significant results (with t value greater than 2.024) were obtained in the following

cases:

Copper tube

Dimensions: outer diameter 35.2mm, inner diameter 32, length 140mm

Following results regard the copper tube into the bore, fixed or in movement (z-transl).

Copper fixed tube bore entrance

Slice n° 1 5 10 20

t SNR 2.334 -0.2844 -0.0649 -0.5053

t SD -0.545 0.0411 -0.6422 -0.3686

Copper tube in movement (z-transl)

Slice n° 1 5 10 20

t SNR 1.1636 0.6517 -0.4203 -1.2636

t SD -0.2386 -0.2449 -0.274 0.3341

Aluminium tube

Dimensions: outer diameter 23.2mm, inner diameter 20, length 250mm.

Following results regard the aluminium tube into the bore, fixed or in movement (z-transl or x-

transl).


11

Aluminium tube fixed bore entrance

Slice n° 1 5 10 20

t SNR 1.6348 -1.9616 -1.7200 0.7207

t SD -0.4615 -0.2243 -0.1466 0.1019

Aluminium tube in movement (z-transl)

Slice n° 1 5 10 20

t SNR 0.3155 -1.6846 -0.2238 -1.1829

t SD -0.0745 -0.2647 -0.2989 -0.0223

Aluminium tube in movement (x-transl)

Slice n° 1 5 10 20

t SNR 3.0601 -0.2022 -1.5856 2.1266

t SD 0.2262 0.1240 -0.1184 -0.0077

Brass tube

Dimensions: outer diameter 20.2mm, inner diameter 16, length 105.5mm.

Following results regard the brass tube into the bore, fixed or in movement (z-transl or x-transl).

Brass tube fixed bore entrance

Slice n° 1 5 10 20

t SNR 0.1085 -0.3448 0.4823 1.3954

t SD -0.0606 -0.4061 -0.1016 -0.5743

Brass tube in movement (z-transl)

Slice n° 1 5 10 20

t SNR 0.1755 -0.6812 -0.0245 -07212

t SD -0.0903 0.1819 0.4089 0.0625

Brass tube in movement (x-transl)

Slice n° 1 5 10 20

t SNR 1.0047 -0.2243 -0.8560 -0.7447

t SD -0.0610 -0.1845 -0.2487 -0.3723

The t test shows the compatibility of these three materials with the MR environment, confirming the

previous tests of [3].


12

Ultrasonic Motor

The piezo-motor we used is a Shinsei USR60-E3N (non-magnetic with encoder) [6].

The following results regard the experiment with the piezo motor on with high or low load at

several distances from the bore entrance and motor off in z-translation.

Piezo motor z translation bore entrance

Slice n° 1 5 10 20

t SNR -0.2963 -0.9721 -0.2835 -0.1105

t SD -1.1647 -0.0765 0.2433 0.0453

Piezo motor 50 cm bore entrance

Slice n° 1 5 10 20

t SNR -0.6649 1.8194 1.2143 0.2233

t SD 0.3545 -0.1398 0.2426 0.1668

Piezo motor bed feet 15 sec on – 15 sec off

Slice n° 1 5 10 20

t SNR 1.3636 0.7301 0.0356 2.3461

t SD -0.8989 -0.6253 -0.2339 -0.3833

Piezo motor on bore entrance

Slice n° 1 5 10 20

t SNR 0.1071 0.5632 0.6427 0.8979

t SD -0.7972 -4.8244 -0.3595 0.0356

Piezo motor on bore entrance high load

Slice n° 1 5 10 20

t SNR -0.3799 -0.0254 -1.15 0.0652

t SD -0.5689 0.0053 0.1244 -0.0286

The results show the compatibility of the piezo motor with the MR environment; in fact the t values

are smaller than the critical t value in each case (a t value greater than the critical value occurs in

only one case but it’s not significant because it’s for a border line image where are possible

artefacts due to instability of the magnetization).


13

DC motor

The DC motor we used is a Maxon Re 40 24V [9].

The following results regard the experiment with the DC motor with high or low load and with or

without RF shield in the MR scanner room, at about 3 meters from the bore entrance.

The previous analysis evidenced the presence of artefacts in the images that caused a SNR decrease

[3] ; the results of the new methods confirm this data.

DC motor off

Slice n° 1 5 10 20

t SNR 0.1293 0.9994 0.6059 1.0641

t SD -0.6049 -0.0781 0.0997 -0.4407

DC motor on with low load

Fetta n° 1 5 10 20

t SNR 3.2886 3.5989 3.0607 4.2183

t SD -3.9957 -3.5301 -3.3858 -3.5823

DC motor on with high load

Fetta n° 1 5 10 20

t SNR -3.2284 3.3528 3.7272 3.7317

t SD -20.1791 -14.0852 -20.9847 -18.1793

DC motor on with high load and RF filter

Fetta n° 1 5 10 20

t SNR 0.8382 2.2447 1.8499 3.3204

t SD -2.5062 -7.1358 -3.2805 -3.3176

The DC motor placed about 3 meters from the bore entrance (inside the MR scanner room) causes

highly significant artefacts in the images, as appears from the previous table (in which the t value,

especially for the deviation standard analysis, is much larger than the critical value).

The use of the RF filter seems to improve the compatibility results.

However, it is impossible to use the DC motor in the MR scanner room. This is why in our system

the DC motor is placed in the console room and power is transmitted using a hydrostatic connection

[4][5]. In this way we avoid most compatibility problems.


14

Scanner stability in the same day

The follow tables show sequences acquired during the same day under the same test conditions

(phantom with any device into the scanner: baseline).

First baseline – second baseline

Slice n° 1 5 10 20

t SNR -0.4121 -1.783 -1.635 -1.050

t SD -0.1371 -0.3484 -0.4076 0.183

First baseline – last baseline

Slice n° 1 5 10 20

t SNR -0.3401 0.0470 1.3858 -1.4092

t SD -0.1263 -0.7271 -0.070 -0.2401

The t value is always smaller than the critical t value, which means that there is no significant

difference between two series of images acquired under the same test condition (baseline: phantom

with no devices inside the scanner) and in the same day: this means that there are no statistically

significant artefacts caused by system instability in the same day .

So, if we compare a sequence regards a test with an under test device into the bore with one of these

baseline and we found artefacts, these are caused by the non perfect compatibility of the device.

Scanner stability in several days

The following tables compare results of baseline measurements obtained on several days under the

same test conditions.

Baseline 12/07/2003 – Baseline 10/07/2003

Slice n° 1 5 10 20

t SNR 7.8168 15.1287 14.2822 12.1362

t SD 1.2378 0.8222 0.3028 0.5409


Slice n° 1 5 10 20

t SNR -3.0988 -1.0982 1.5479 4.0722

t SD -0.3153 0.0027 0.3682 0.1614


Slice n° 1 5 10 20

t SNR -10.1395 -14.388 -10.5120 -8.6411

t SD -0.8449 -0.7125 -0.2443 -0.3806


15

There are significant differences between baselines acquired in several days. Therefore the device

analysis must be performed relative to the baseline of the corresponding day.

Shielded cable and linear potentiometer

The following results refers to the experiment with the electric cable that we tested with or without

linear potentiometer in two cases: current flow and no current flow through the cable at several

distances from the bore and with or without RF shield.

Shielded cable with no current Slice n° 1 5 10 20

t SNR 0.4927 -1.2331 -0.1442 -1.6831

t SD 0.2585 0.1568 0.1539 -0.2677

Shielded cable with linear potentiometer and no current Slice n° 1 5 10 20

t SNR 1.0668 2.9799 2.4716 2.8136

t SD -1.8813 -1.4334 -1.3499 -1.6861

Shielded cable with linear potentiometer and current Slice n° 1 5 10 20

t SNR 1.3769 2.9576 2.2219 0.2269

t SD -0.8236 -0.9341 -1.0097 -1.1117

Shielded cable with no current 40 cm inside the bore Slice n° 1 5 10 20

t SNR 0.5827 1.0836 0.6218 1.2294

t SD -0.5548 -0.1926 -0.4883 -0.6074

Shielded cable with linear potentiometer and current at 80 cm from the bore Slice n° 1 5 10 20

t SNR 4.7915 5.111 5.5723 3.1689

t SD -2.1682 -2.1252 -2.1469 -2.4087

Shielded cable with linear potentiometer, current and RF filter at the bore entrance Slice n° 1 5 10 20

t SNR 0.2782 2.7140 2.1476 0.3566

t SD -0.6390 -0.8781 -1.1157 -1.1544

Shielded cable with linear potentiometer and current at the bore entrance

Slice n° 1 5 10 20

t SNR 0.2599 2.2259 1.8817 1.6065

t SD -0.7341 -0.5845 -1.1468 -1.2426


16

The previous tables show that there are some problems when the linear potentiometer is connected

to the shielded cable with the current flows through it. This problem may be due to the last piece of

shielded cable, constituted by two not shielded wires that may create artefacts on the image.

In future tests we will test the linear potentiometer with particular attention to the connection with

the power cable.

The use of the RF filter seems to minimize these artefacts (the difference is evident analyzing the t

values regarding the SD analysis).

3.2 Haptic interface compatibility test In the final parts of this work we tested the complete prototype of the haptic interface in the MR

environment in order to evaluate its compatibility such as the absence of artefacts in the images.

We used the methods described in the section 3.1: we acquired the images using a spherical

phantom and a GE-EPI (gradient echo, eco planar imaging) sequence in the following cases:

- baseline: phantom with no devices inside the scanner

- device off ( see Figure 2 B ) inside the bore

- device inside the bore, rotation parts on (see Figure 2 A) (U-motor on, hydraulics part and

DC motor off)

- device inside the bore, rotation and translation parts on (U motor on, hydraulics part and DC

motor on) with sensor part on (see Figure 1 ): complete operation mode.

The hydraulics part and the master rack (electronic driver and the emergency button, see Figure 3)

were positioned in the console room and the electric cables and the hydraulic ones passed through

the wires guides in the Faraday’s shield. The electric cables are shielded and, moreover, we used a

RF filters on the plug- in with the master rack.

3.2.1 Results

The following tables show the results obtained by the statistical test used for last analysis (t test) for

the cases described previously; there are statistically significant results with a t value greater than

2.024.

Baseline: scanner stability in the same day

We acquired several baseline images in order to evaluate the scanner temporal stability: the

following tables report the t values regard the differences between different baselines acquired in

the same day but in different moments.


17

First baseline – second baseline

Slice n° 1 5 10 20

t SNR -0.8060 0.6737 -0.7408 -1.6288

t SD -0.2332 0.2367 0.9364 0.1043

First baseline – last baseline

Slice n° 1 5 10 20

t SNR 1.2075 -0.3686 -1.1303 0.2727

t SD -0.3681 -0.0361 0.2637 -0.2083

It’s possible to note that in the same testing session there aren’t problem due to the scanner

environment stability.

Baseline: scanner stability in several days

The following tables compare results of baseline measurements obtained on several days under the

same test conditions.

First baseline 02/04/2004 – First baseline 03/04/2004

Slice n° 1 5 10 20

t SNR 1.4001 -0.9679 0.2825 -8.4784

t SD -0.3090 -0.0667 0.6230 0.3312

First baseline 02/04/2004 – Last baseline 03/04/2004

Slice n° 1 5 10 20

t SNR -2.9608 -1.5047 1.3336 -3.4352

t SD 0.6711 -0.1039 0.6927 0.0626

First baseline 02/04/2004 – Last baseline 04/04/2004

Slice n° 1 5 10 20

t SNR -0.4946 0.7838 2.0114 -3.7732

t SD 0.3662 -0.1074 0.6634 0.4013

In this new test session there weren’t artifacts due to MR environment instability: in fact statistical t

test for differences between baselines acquired in different days, don’t show a t value greater than

the critical t value (a t value greater than the critical value occurs in some case but it’s not

significant because it’s for border line images where are possible artefacts due to instability of the

magnetization).


18

Device off inside the bore

Device was positioned inside the bore in the right position for the experiment with a subject (near to

his right hand); power was off both for the two motors and for the sensor part.

Respect to the first baseline

Slice n° 1 5 10 20

t SNR 0.5100 1.1159 0.8301 -1.3576

t SD 0.0550 0.1766 0.4974 -0.3034

Respect to the last baseline

Slice n° 1 5 10 20

t SNR 1.5131 1.3395 0.6156 -0.4374

t SD 0.5715 0.4300 0.4326 -0.0107

The previous results show the good MR compatibility of the parts of the haptic interface that must

be positioned inside the bore in order to stimulate the subject fingers.

Device inside the bore, rotation part on (U-motor on, hydraulic part and DC motor off)

In this experiment we tested the rotation part of the device: the stimulation pad rotates by mean the

ultrasonic motor inside the scanner. Rotation direction, velocity and temporization are controlled by

a PC with an acquisition card placed inside the console room.

Rotation (sequence: 15 sec on 15 sec off) respect to the first baseline

Slice n° 1 5 10 20

t SNR 2.2342 -0.2353 -0.7089 0.4743

t SD -0.4409 -0.0956 0.5130 -0.2278

Rotation (sequence: 15 sec on 15 sec off) respect to the last baseline

Slice n° 1 5 10 20

t SNR 0.9752 1.4692 -0.8140 -0.4374

t SD 0.5715 0.4300 0.4326 -0.0107

Rotation (sequence: 15 sec on 15 sec off) respect to the first baseline of the last testing day

Slice n° 1 5 10 20

t SNR 1.5622 -1.0791 0.9533 0.7699

t SD 0.3245 -0.4133 0.1965 0.3591

The device rotation don’t act any artefact in images: in fact the good compatibility of the U-motor

was already demonstrated in [3] and in the section 3.1.1 of this report.


19

Device inside the bore, complete operation mode

Following tables regard the tests of the device complete operation mode: translation and rotation

parts ON and sensor part (linear potentiometer) ON.

These experiments were performed on a different day respect the previous so the reference series

for the statistical t test are the baselines acquired in that day.

Ultrasonic motor and DC motor ON; potentiometer plug off

Slice n° 1 5 10 20

t SNR 1.1846 -1.0944 0.7291 -1.1248

t SD -0.0789 0.5738 0.2640 0.4482

Ultrasonic motor and DC motor ON (seq: 15 sec on 15 sec off); potentiometer plug off

Slice n° 1 5 10 20

t SNR 2.1843 -0.4931 -0.1762 -0.0538

t SD 0.2312 -0.5531 0.2769 -0.1559

Ultrasonic motor and DC motor ON; potentiometer plug ON

Slice n° 1 5 10 20

t SNR 0.2795 1.1197 -0.0256 -1.0419

t SD 0.4073 0.0864 0.1243 -0.1308

Ultrasonic motor and DC motor ON (seq: 15 sec on 15 sec off); potentiometer plug ON

Slice n° 1 5 10 20

t SNR 1.8152 0.1332 -1.0693 1.3127

t SD -0.2293 -0.1513 -0.1048 -0.0719

During this test session the haptic interface worked completely like as it must be work during a

stimulation experiment on a subject: the previous results show the good compatibility of the device

(a t value greater than the critical value occurs in only one case but it’s not significant because it’s

for a border line images where are possible artefacts due to instability of the magnetization).

3.3 Discussion

Results for aluminium, copper and brass indicate no differences with reference images acquired in

the same day, both for SNR and SD values. Experiments with these materials moved by an operator

led to the same conclusions. Statistical differences were found for the electric cable with the

potentiometer plugged both with current flowing (0.25 mA) and with no current. Experiments were

performed with motors in two conditions: turned alternatively on and off for 15 seconds intervals,


20

and always off. The ultrasonic motor showed no differences with reference images in both

conditions while the DC motor showed significant differences, even if placed in the most distant

corner of the MRI room, at about 3 meters from the scanner bore. Figure 4 shows the homogeneous

phantom image (baseline) and result of the difference between the base image and the image

acquired during the DC motor running. Figure 5 shows the relation between SNR and time during

the DC motor running (sequence: 15 s ON, 15 s OFF).

Figure 4 Homogeneous phantom image (baseline) (left) and result of the difference between the base image and the image acquired during the DC motor running (right).

Figure 5 Relation between SNR and time (continuous line) during the DC motor running (outlined line: motor ON=high level, motor OFF=low level).


21

Regarding the temporal stability of scanner measurements, the results indicated that reference

images acquired in the same day showed no statistically significant differences. Images in the same

conditions but acquired on different days showed a statistical difference in SNR values, with no

difference for standard deviations.

Tests regard the complete haptic interface show the good compatibility of this device: the device off

was positioned inside the scanner room on the patient bed in the right position for the stimulation

experiment with a subject. The images acquired in these experimental conditions don’t show

statistical differences with the reference images (baseline: phantom with no device).

The ultrasonic motor is placed inside the scanner on the stimulation part with the linear

potentiometer: power and sensor signals are sent from the console room to the scanner room by

mean shielded cable through the wires guides in the Faraday’s shield. The images acquired during

the rotation movement of the stimulation pad don’t show artifacts so this part of the interface can be

considered MR compatible.

Finally we tested the complete function of the haptic interface: rotation, translation and sensor part

with both motors ON and the linear potentiometer power ON for the pad position detection.

Also in this case the statistical test didn’t show statistical differences between the reference images

and the images acquired in the experimental conditions of subject stimulation so it’s possible to

conclude that the designed haptic interface has a good compatibility with the MR environment and

it’ possible to use it for functional experiment of a volunteer.

In Figure 6 it’s possible to see the result of the difference between two baselines acquired in the

same day but in two different moments and result of the difference between the first baseline and

the image acquired during the haptic interface complete operation mode.

Figure 6 Result of the difference between two baselines acquired in the same day but in two different moments (left) and result of the difference between the first baseline and the image acquired during the haptic interface complete

operation mode (right).


22

4 Conclusions We have developed an MR compatible haptic interface with two degrees of freedom to investigate

the brain mechanisms of tactile perception. The interface uses active as well as passive MR

compatible actuators and sensors. In order to study the eventual presence of artifacts in the MR

images due to the device function, a statistical t test was used. In this report results of compatibility

experiments of this interface are reported.

While this prototype concludes the contribution of EPFL to Touch-HapSys, EPFL will continue

collaboration with UNIPI to adapt control of the interface to different studies, as well as to develop

a force sensor to measure direction and amplitude of finger friction force exerted against the haptic

interface and carry out behavioural and fMRI studies using this device in collaboration with UNIPI.

5 References [1] Chinzei, K., Kikinis, R., Jolesz, F.A., MR Compatibility of Mechatronic Devices: Design

Criteria, Proc MICCAI ’99, Lecture Notes in Computer Science, 1999, 1679: 1020-31.

[2] Chinzei K., Miller K., MRI Guided Surgical Robot, Proc. 2001 Australian Conference on

Robotics and Automation, Sidney 14-15 November 2001.

[3] V. Hartwig, N. Vanello, N. Sgambelluri, E. Scilingo, L. Landini, A. Bicchi, D 4.10: Design of

fMRI compatible actuators, http://www.touch-hapsys.org/, September 2003.

[4] R. Moser, R. Gassert, E. Burdet, L. Sache, H. R. Woodtli, J. Erni, W. Maeder and H. Bleuler,

An MR Compatible Robot Technology, Proc. IEEE International Conference on Robotics and

Automation (ICRA), 2003.

[5] G. Ganesh, R. Gassert, E. Burdet and H. Bleuler, Dynamics and Control of an MRI Compatible

Master-Slave System with Hydrostatic Transmission, Proc. IEEE International Conference on

Robotics and Automation (ICRA), 2004

[6] Shinsei Corporation Inc., Ultrasonic motor USR60 series,

http://www.tky.3web.ne.jp/~usrmotor/English/html/


23

[7] GE Medical System (ed): MR Safety and MR Compatibility: Test Guidelines for Signa SP TM

Version 1.0, http:// www.gemedicalsystems.com/rad/mri/pdf/safety1.pdf , October 1997.

[8] V. Hartwig, N. Vanello, R. Gassert, D. Chapuis, M.F. Santarelli, V. Positano, E. Ricciardi, P.

Pietrini, L. Landini, A. Bicchi: A compatibility test for tactile displays designed for fMRI

studies, accepted to EuroHaptics 2004, Munich, June 2004.

[9] Maxon DC motor Data Sheet, RE 40 diameter 40 mm, Graphite Brushes, 150 Watt,

http://www.maxonmotorusa.com/files/catalog/2003/pdf/03_082_e.pdf

Documents

Prototype fMRI compatible actuators Final version€¦ · Moreover we tested the final prototype of the haptic interface to analyze its compatibility with the MR environment and we