CONSISTENT CORTICAL RESPONSES FROM SUBCORTICALY

DELIVERED ELECTRICAL STIMULI

A Study Oriented to Visual Prostheses

Fivos Panetsos, Elena Diaz-de Cerio, Abel Sanchez-Jimenez, Juan Jose Navarro-Valls

Neurocomputing and Neuro-robotics Research Group, and School of Optics, Complutense University of Madrid

Avda Arcos de Jalon s/n, Madrid, Spain

Jose A. Vega

Department of Anatomy and Embryology, Faculty of Medicine, University of Oviedo, Oviedo, Spain

Idoia Diaz-Guemes

Center of Minimal Invasion Surgery “Jesus Uson”, Caceres, Spain

Keywords: Visual prostheses, neuroprostheses, lateral geniculate nucleus, thalamus and implants.

Abstract: Loss of vision is one of the most important challenges for science nowadays and a large amount of work has

been done in the development and implant of visual neuroprostheses. The applicability of retinal implants is

restricted either because healthy retinal neurons and/or optic nerve are not always available or because of

problems related to the retinal implants themselves. At the present alternatives are restricted to cortical

prostheses which in turn have several physiological and technical limitations. In our communication we

describe a direct proof for the feasibility of subcortical visual prostheses that would solve several of the

limitations of the cortical ones. Our approach consists in stimulating the visual cortex of intact animals by

means of visual stimuli and then to generate similar responses by means of electrical stimulation of the

lateral geniculate nucleus.

1 INTRODUCTION

1.1 General Context

Loss of vision is one of the most important

challenges for science nowadays. Vision impaired

people are among the most vulnerable and

emarginated, normally with low incomes or

unemployed, with poor education, inadequate social

protection, many problems for public or private

transport, perception of the environment, access to

the buildings, etc. In addition to these adversities,

the negative attitudes of the society create a hostile

environment to both, the blind people and their

families.

For all these reasons the development of visual

prostheses represents one of the highest priorities in

the field of Biomedical Engineering, despite vision

is the most complicated of our senses and several

extremely complicated physiological, computational

and engineering problems have to be solved at every

step. Up to now, a large amount of work has been

done in the development and chronic implants of

visual neuroprostheses (for a thorough review see

(Maynard, 2001)) including chronic implants in the

retina, optic nerve and occipital cortex of blind

human subjects (Cohen, 2007; Dobelle, 2000;

Gerding, 2007; Humayun et al., 2003; Javaheri et al.,

2006; Lakhanpal et al., 2003; Margalit et al., 2002;

Shenoy et al., 2006; Thanos et al., 2007; Veraart et

al., 1998; Winter et al., 2007).

Retinal prostheses are very useful for the

impaired people but they need functional retinal

neurons and intact optic nerve, thalamus and cortex.

On the other hand cortical prostheses have to be

directly interfaced and inject visual signals to a high-

level neural structure, the visual cortex, designed by

nature to receive and analyse complex information.

155

Panetsos F., Diaz-de Cerio E., Sanchez-Jimenez A., Navarro-Valls J., Vega J. and Diaz-Guemes I. (2009).

CONSISTENT CORTICAL RESPONSES FROM SUBCORTICALY DELIVERED ELECTRICAL STIMULI - A Study Oriented to Visual Prostheses.

In Proceedings of the International Conference on Health Informatics, pages 155-160

DOI: 10.5220/0001510201550160

 SciTePress

Moreover, such information is previously pre-

processed by the retina and visual thalamus and this

process includes an open loop formed by the

corticothalamic and thalamocortical projections

(Mason et al., 1991) the function of which is not

taken into account in the design of cortical

prostheses.

Most of the problems related to the architecture

of the visual cortex and the input of visual signals

could be solved if visual prostheses would be

implanted to the previous relay station of the visual

system, the lateral geniculate nucleus of the

thalamus; such prostheses would also benefit from

the processing capabilities of the thalamocortical

loop (see Section 1.2). But to our knowledge there is

not available experimental work on visual prostheses

to implant in subcortical structures and in particular

in the visual thalamus, except the work of Pezaris

and Reid (2007) in the generation of visual percepts

after stimulation of the lateral geniculate nucleus.

In the present paper we describe our work on

visual prostheses implanted in the lateral geniculate

nucleus of the thalamus and a direct way to

demonstrate the appearance of artificially generated

cortical responses similar to those elicited by natural

visual stimuli.

1.2 The Visual System

The visual system processes information in a

hierarchical manner from the retina to the cortex

through increasing the complexity of feature

extractions using a chain of three neurons and two

neural relay stations connected in a massively

parallel fashion. The first information processing

station is the eyeball’s inner lining, the retina, where

a bidimensional sheet of photoreceptors transform

the image of the external world to a multi-

dimensional spatiotemporal and intensity pattern of

electrical signals (Baylor et al., 1979; Saito et al.,

1978) further processed by the bipolar, horizontal,

amacrine and ganglion cells (Kuffler, 1953). Via the

optic nerve, the outputs of the two retinas are

transmitted to the lateral geniculate nucleus of the

thalamus (LGN) where are processed (Kastner et al.,

2006; Derrington and Fuchs, 1979) and integrated

into a binocular representation of the world (Murphy

and Sillito, 1989).

LGN output goes directly to the primary visual

cortex (V1) which gives in turn a massive feedback

to LGN and also sends information to higher visual

cortical regions for further processing (Mason et al.,

1991). The corticothalamic information feedback

plays a major role to information processing of the

visual signal, selecting the most interesting inputs

and imposing processing rules to the thalamus

(Sillito et al., 2006).

1.3 Visual Prostheses

Blindness results from either an inability of the

visual system to transduce light energy into electric

signals or a failure of the generated electrical signals

to reach the higher relay stations of the visual

pathway. It is classically accepted that electrical

stimulation of V1 can elicit a visual percept of light

denominated phosphene (Tehovnik and Slocum,

2007; Krisch and Hosticka, 2007).

In retinal prostheses arrays of electrodes are

placed either on the retinal surface or in the

subretinal space where they stimulate ganglion cells

(Humayun, 2003). A new approach to retinal

prosthesis is electrical stimulation of the retina with

extraocular electrodes (Chowdhury et al., 2008).

However they need an undamaged retina and an

intact optic nerve (Margalit, 2002).

Cortical prostheses consist of arrays of electrodes

which are placed to or penetrate the cortical surface

and stimulate layer VI of the visual cortex to create

discrete phosphenes (Brindley and Lewin, 1968;

Dobelle, 2000) and have the advantage to be suitable

for almost any kind of blind patients.

However the visual cortex is non-linear and non-

conformal with visual space, not letting us to predict

in a precise way where will phosphenes be elicited

when stimulating with each electrode (Warren et al.,

2001). Moreover, the power of cortical processing is

mainly based to the continuous feedback of the

thalamocortical loop and to the influence V1

exercises to the thalamus, both of them excluded due

to the direct introduction of visual information to

V1.

Figure 1: Schematic representation of the visual pathway.

In the cases in which the retina and/or the optic

nerve are damaged or not functioning, the target

with more advantages and less technical,

experimental or clinical problems for visual

prostheses seems to be the thalamus:

HEALTHINF 2009 - International Conference on Health Informatics

156

1) the receptive fields (RFs) of LGN neurons are

simple, well characterized, and similar to those of

their retinal afferents (Hubel and Wiesel, 1961;

Wiesel and Hubel, 1966)

2) fovea and parafovea are spatially represented

in the LGN facilitating the accessibility of neurons

with central visual fields (Pezaris and Reid, 2007)

3) LGN cells give rise to axons that terminate in

primary visual cortex, in a highly specific manner

making monosynaptic connections with simple cells

predominantly when the pre- and postsynaptic

receptive fields overlap and match in sign, size, and

time course (Alonso et al., 2001), consequently

stimulation of a small number of LGN neurons

should achieve simple, focal percepts (Pezaris and

Reid, 2007)

4) LGN cells receive a massive cortical feedback

that directly control the processing capabilities of

these neurons conditioning and selecting the visual

information from LGN to the cortex (Sillito et al.,

2006). Consequently, LGN implants will use the

plasticity and adaptability of the Central Nervous

System to modify the responses of its own neurons

to recognize the artificial signals through the

influence of the corticothalamic LGN-V1 loop.

Despite the above considerations LGN has never

been taken into account in the development of

prosthetic visual devices, except the mentions in

Pezaris and Reid (2007).

The direct test of the feasibility of visual devices

implanted in the thalamus is just the target of the

present work. The data we present here come out

from our research carried out within the framework

of our projects on visual neuroprostheses. In this

work our objective was to generate perception

sensations similar to the natural ones and assess

them by comparing the responses of the cortical

neurons to electrical stimulation of the thalamus

with those generated by visual stimulation of the eye

(see Section of Materials and Methods).

2 EXPERIMENTS

2.1 General Approach

Data were obtained from 36 urethane- (1.5g/kg i.p.)

or sodium pentobarbital (35mg/Kg i.p.)

anaesthetized Wistar rats of both sexes, weighting

200-240g. Experiments were carried out according

to the national legislation (R.D. 1201/2005) and EU

Directives on this matter (86/609/EC). Rats have

been used due to their simple but complete visual

system. In a previous work we had realised a

mapping of the thalamic and cortical visual areas

and to characterized the responses of their neurons

to simple and complex visual stimuli. The 36

animals were used in a series of combined

experiments in which one group of electrodes was

placed in the LGN and another group in V1.

Our approach consisted in: 1) presenting a series

of visual stimuli to one eye and record the responses

of the contra lateral LGN and V1, R

and R

respectively, by means of the implanted

multielectrodes, 2) inject to LGN an electrical

pattern R

similar to the previously recorded R

during the presentation of the visual stimuli; at the

same time record the cortical responses R

to this

electrical stimulation of LGN and 3) modify the

parameters of R

looking for the best matching

between R

and R

, that means between the

response to natural and the response to the electrical

stimuli.

The ability to elicit R

responses similar to R

for a large number of natural visual stimuli would be

a proof of the feasibility of visual prostheses

implanted to the thalamus.

In total, 216 complete cycles of experiments

(visual stimulation – LGN & V1 recordings,

electrical stimulation – V1 recordings were

performed.

2.2 Stimulation and Recordings

Animals were placed in a stereotaxic device that

enables the conduct of visual experiments. An

incision was performed and two holes were made in

the skull to allow access to the rat brain in the

appropriate coordinates. Recording and stimulating

multielectrodes were developed and tested in our lab

following the methodology described by Neuralynx

(http://www.neuralynx.com) and then introduced in

the brain. Anaesthesia level was controlled by the

amplitude of the EEG waves. A frontal hole was

made to record the electroencephalogram (EEG).

EEG recordings were performed through an

insulated (except in the tip) 1mm diameter Cr-Ni

macroelectrode introduced in the frontal cortex at

1.0mm from the surface. The EEG was continuously

monitored in the oscilloscope. In case of reduction

of the amplitude of the waves supplementary doses

of anaesthesia were administered.

2.3 Data Acquisition and Analysis

Single channel recordings were performed using

tungsten microelectrodes (2.0MΩ) and Micro1401

hardware by Cambridge Electronic Design with

CONSISTENT CORTICAL RESPONSES FROM SUBCORTICALY DELIVERED ELECTRICAL STIMULI - A Study

Oriented to Visual Prostheses

157

accompanying software Spike2. Multichannel

recordings were performed using the above

described multielectrodes and neural activity was

acquired using a PCI-6071E E Series data

acquisition card from National Instruments, with

accompanying Recorder software amplified and

displayed on a Plexon Inc PCI device, stored and

then imported to a Spike2 software and analyzed

using MATLAB (©MathWork corporation) and

Spike2 software. Data were sampled and digitalized

at 20 KHz, stored in personal computers and then

processed off-line. We first performed single

channel recordings both in LGN and V1 with

tungsten electrodes in order localize a region with

response to visual stimulation.

Once the most suitable region was identified the

tungsten electrode was substituted by a 4x4

multielectrode array. The neural tissue was then let

to recover its normal activity for 20 minutes and

then we started with the standard

recording/stimulation procedure. The exact location

of the recording electrode was also confirmed on

subsequent histological preparations.

Spikes were threshold-isolated offline using

Spike2 software (Cambridge Electronic Design)

taking as threshold a value equal at least three times

the level of the noise, and converted into discrete

processes. To determine basic features of neuronal

response and behaviour we performed peristimulus

histograms, interspike interval histograms, auto- and

cross-correlation histograms.

2.4 Histology

To ascertain the localization of the electrodes, the

histology of the brain was analyzed 1 mm rostral and

dorsal to the electrodes placed in the LGN and the

occipital cortex, respectively.

Briefly, after recording, and because the

electrodes are too thick to easily identify the cerebral

structured where they are placed into, a 2mA electric

current was passed though them for 10s in order to

electro coagulate the recorded structures. Then the

animals were sacrificed with an intracardiac

injection of NaCl hypertonic solution (3ml) and the

brain was quickly removed and whole fixed in

Bouin’s fixative for 24 hours, then dehydrated

through increased concentrations of ethanol (from

70º to absolute) and xylene to remove the picric

acid, and embedded in paraffin.

The blocks were cut serially to obtain coronal

sections 10μm thick, deparaffinized, rehydrated and

stained with methylene blue-eosin. The sections

were then washed in tap-water, dehydrated and

mounted with Entellan®, and studied in a light

photomicroscope. The electrode implanted into the

lateral geniculate body has a trajectory perpendicular

to the brain surface. The structural techniques used

to identify the lesioned zones were routinely

haematoxylin & eosin and alzian blue &

haematoxylin. Both techniques consented identify

the structure of the brain and the electrocoagulation-

induced lesion.

2.5 Stimulations and Responses

In a first phase, visual stimuli of increasing

complexity were used to determine the response

patterns evoked in both LGN and V1. Two types of

visual stimuli were used, flashes (Grass model PS33,

20-40 stimuli at 1Hz) and persistent geometric black

and white figures (see figure 2): horizontal stripes,

circle, ring of light, cross and black cross on a white

background. Persistent stimuli were generated on the

screen of a PC (324 X 244 mm and 1024 x 768

pixels resolution) applied for 3 seconds at 0.3Hz.

Every stimulus had a TTL synchronization signal

toward the data acquisition system.

Figure 2: Geometric stimuli used in the experiments.

Responses were analyzed, correlated and patterns

of electrical stimuli were generated and applied to

the LGN according to the procedure described in

Section 2.2. Then it has been possible to extract

basic characteristics of the electrical stimulation that

evoked cortical responses comparable to those of the

visual stimuli.

After that, each animal was applied a battery of

visual stimuli, thalamic and cortical activity were

recorded, online analyzed, and sets of electrical

stimulation patterns were generated. Such

stimulation patterns were then applied to the LGN,

and cortical responses were recorded and compared

to those generated by the visual stimuli (see figure

3).

HEALTHINF 2009 - International Conference on Health Informatics

158

Current was applied across the four electrodes in

the geniculate nucleus using SIUs (World Precision

Instruments, A365 and A360 Stimulus Isolator Unit)

controlled by the Spike software with the Micro1401

mkII data acquisition unit (Cambridge Electronic

Design). We used intensities between 100μA and

600μA and applied different patterns of stimulation

differing in the number of stimuli trains, the interval

between trains, the number of stimuli per train and

their duration and the number of electrodes and the

temporal relation between the current applied by

each electrode.

Figure 3: A visual stimulus is presented to the eye and

neural responses are recorded from the lateral geniculate

nucleus in the thalamus and the primary visual cortex (up).

The characteristics of the thalamic electric activity are

extracted and an artificial electrical stimulus is generated

and delivered to the thalamic neurons in absence of

external (natural) visual stimulus (down). A fine

adjustment of the spatiotemporal and intensity

characteristics of the artificial stimuli delivered through

the multiple electrodes implanted to the thalamus allow us

to obtain responses from the cortical neurons that are very

similar to those induced by the natural stimulation using

geometric stimuli. Similar results were obtained from the

entire set of visual stimuli (geometric forms) and the

corresponding electrical stimuli.

3 CONCLUSIONS

Due to the architectonic organisation of the lateral

geniculate nucleus and the lower complexity of the

processing of visual information it performs (if

compared to the visual cortex) LGN is the best

candidate for visual implants when the retina and/or

optic nerve are not functional. Moreover, the direct

action of the visual cortex to the thalamic neurons

through the thalamocortical loop allows a better

adaptation of the Central Nervous System to the

artificial input to the thalamus than to the cortex.

In the present paper we prove that electrical

stimulation of the lateral geniculate nucleus can

generate neural responses in the visual cortex that

resemble those elicited by natural visual stimuli.

Such responses can be achieved after a sampling of

the thalamic responses to the natural stimuli by

means of multielectrode recordings, the extraction of

their basic spatiotemporal characteristics and a

subsequent fine tuning of the electrical stimuli

delivered through the same electrodes implanted into

the thalamus.

Our results are important for the development of

visual prostheses implanted in the subcortical

structures of the brain of blind people although an

extensive work has to be done: in addition to

anatomophysiological problems related to the

implants of the electrode, damages to the brain due

to the chronic stimulation, etc., research on coding

complex visual stimuli, images in movement,

reduction of the complexity of the visual image. So

future steps to solve these problem will be recordand

stimulate with more channels (up to 100), applied

visual stimuli in movement, implant multielectrodes

chronically, mathematical study of the interactions

between recording/stimulating channels to develop

more precise microstimulation, etc.

ACKNOWLEDGEMENTS

Supported by a grant of MAPFRE-Medicine

Foundation (“Connections between Central Nervous

System and electronic devices”, 2004), a grant of

MAPFRE Foundation (“Visual Prostheses”, 2005)

and a grant of the Spanish National Organisation of

Blinds O.N.C.E. (VISNE project: “VISual

neuroprostheses based on adaptive NEuron-silicon

interfaces”, 2006-2009).

REFERENCES

Alonso, J.M., Usrey, W.M., & Reid, R.C. (2001). Rules of

connectivity between geniculate cells and simple cells

in cat primary visual cortex. Journal of Neuroscience,

21, 4002-4015.

Baylor, D.A., Lamb, T.D., & Yau, K.W. (1979).

Responses of retinal rods to single photons. J.Physiol,

288, 613-634.

Brindley, G.S. & Lewin, W.S. (1968). The sensations

produced by electrical stimulation of the visual cortex.

J.Physiol, 196, 479-493.

CONSISTENT CORTICAL RESPONSES FROM SUBCORTICALY DELIVERED ELECTRICAL STIMULI - A Study

Oriented to Visual Prostheses

159

Chowdhury, V., Morley, J.W. & Coroneo, M.T. (2008).

Development of an extraocular retinal prosthesis:

Evaluation of stimulation parameters in the cat.

J.Clin.Neurosci. 15, 900-906.

Cohen, E.D. (2007). Prosthetic interfaces with the visual

system: biological issues. J.Neural Eng, 4, R14-R31.

Derrington, A.M. & Fuchs, A.F. (1979). Spatial and

temporal properties of X and Y cells in the cat lateral

geniculate nucleus. J.Physiol, 293, 347-364.

Dobelle, W.H. (2000). Artificial vision for the blind by

connecting a television camera to the visual cortex.

ASAIO Journal, 46, 3-9.

Gerding, H. (2007). A new approach towards a minimal

invasive retina implant. J.Neural Eng, 4, S30-S37.

Hubel, D.H. & Wiesel, T.N. (1961). Integrative action in

the cat's lateral geniculate body. J.Physiol, 155, 385-

398.

Humayun, M.S., Weiland, J.D., Fujii, G.Y., Greenberg, R.,

Williamson, R., Little, J., Mech, B., Cimmarusti, V.,

Van, B.G., Dagnelie, G., & de, J.E. (2003). Visual

perception in a blind subject with a chronic

microelectronic retinal prosthesis. Vision Research,

43, 2573-2581.

Javaheri, M., Hahn, D.S., Lakhanpal, R.R., Weiland, J.D.,

& Humayun, M.S. (2006). Retinal prostheses for the

blind. Annals of the Academy of Medicine, Singapore,

35, 137-144.

Kastner, S., Schneider, K.A. & Wunderlich, K. (2006).

Beyond a relay nucleus: neuroimaging views on the

human LGN. Prog.Brain.Res 155, 125-143.

Krisch, I. & Hosticka, B.J. (2007). Restoring visual

perception using microsystem technologies:

engineering and manufacturing perspectives. Acta

Neurochir.Suppl. 97, 473-480.

Kuffler, S.W. (1953). Discharge patterns and functional

organization of mammalian retina. Journal of

Neurophysiology, 16, 37-68.

Lakhanpal, R.R., Yanai, D., Weiland, J.D., Fujii, G.Y.,

Caffey, S., Greenberg, R.J., de Juan E Jr, & Humayun,

M.S. (2003). Advances in the development of visual

prostheses. Current Opinion in Ophthalmology, 14,

122-127.

Margalit, E., Maia, M., Weiland, J.D., Greenberg, R.J.,

Fujii, G.Y., Torres, G., Piyathaisere, D.V., O'Hearn,

T.M., Liu, W., Lazzi, G., Dagnelie, G., Scribner, D.A.,

de Juan E Jr, & Humayun, M.S. (2002). Retinal

prosthesis for the blind. Survey of Ophthalmology, 47,

335-356.

Mason, A., Nicoll, A., & Stratford, K. (1991). Synaptic

transmission between individual pyramidal neurons of

the rat visual cortex in vitro. Journal of Neuroscience,

11, 72-84.

Maynard, E.M. (2001). Visual prostheses.

Annu.Rev.Biomed.Eng, 3, 145-168.

Murphy, P.C. & Sillito, A.M. (1989). The binocular input

to cells in the feline dorsal lateral geniculate nucleus

(dLGN). J.Physiol, 415, 393-408.

Pezaris, J.S. & Reid, R.C. (2007). Demonstration of

artificial visual percepts generated through thalamic

microstimulation. Proc.Natl.Acad.Sci.U.S.A, 104,

7670-7675.

Saito, T., Kondo, H., & Toyoda, J. (1978). Rod and cone

signals in the on-center bipolar cell: their different

ionic mechanisms. Vision Research, 18, 591-595.

Shenoy, K.V., Santhanam, G., Ryu, S.I., Afshar, A., Yu,

B.M., Gilja, V., Linderman, M.D., Kalmar, R.S.,

Cunningham, J.P., Kemere, C.T., Batista, A.P.,

Churchland, M.M., & Meng, T.H. (2006). Increasing

the performance of cortically-controlled prostheses.

Conf.Proc.IEEE Eng Med.Biol.Soc., Suppl, 6652-

6656.

Sillito, A.M., Cudeiro, J., & Jones, H.E. (2006). Always

returning: feedback and sensory processing in visual

cortex and thalamus. Trends in Neurosciences, 29,

307-316.

Tehovnik, E.J. & Slocum, W.M. (2007). Phosphene

induction by microstimulation of macaque V1. Brain

Res.Rev. 53, 337-343.

Thanos, S., Heiduschka, P., & Stupp, T. (2007).

Implantable visual prostheses. Acta Neurochir.Suppl,

97, 465-472.

Veraart, C., Raftopoulos, C., Mortimer, J.T., Delbeke, J.,

Pins, D., Michaux, G., Vanlierde, A., Parrini, S., &

Wanet-Defalque, M.C. (1998). Visual sensations

produced by optic nerve stimulation using an

implanted self-sizing spiral cuff electrode. Brain

Research, 813, 181-186.

Warren, D.J., Fernandez, E., & Normann, R.A. (2001).

High-resolution two-dimensional spatial mapping of

cat striate cortex using a 100-microelectrode array.

Neuroscience, 105, 19-31.

Wiesel, T.N. & Hubel, D.H. (1966). Spatial and chromatic

interactions in the lateral geniculate body of the rhesus

monkey. Journal of Neurophysiology, 29, 1115-1156.

Winter, J.O., Cogan, S.F., & Rizzo, J.F., III (2007).

Retinal prostheses: current challenges and future

outlook. J.Biomater.Sci.Polym.Ed, 18, 1031-1055.

HEALTHINF 2009 - International Conference on Health Informatics

160