THE CONTROL OF LOWER LIMB SENSORY INPUTS ASSOCIATED WITH A

SEATED INVERTED-PENDULUM BALANCING TASK

A Thesis

Presented to The Faculty of Graduate Studies

of The University ofGuelph

by DARCY CLAIR BISHOP

In partial fûifihent of requkrnents for the degree of Master of Science August, 1999

@ Darcy Clair Bishop, 1999

National Library

Bibliothèque nationale

du Canada

Acquisitions and Bibliographk SeMces

Acquisitions et services bibliographiques

395 Weiisngton Street ûttawaON K1AON4

395, rue Wdlingtori OrtawaON K t A W Canada

Canada

The author has granted a nonexclusive licence allowing the National Library of Canada to reproduce, lom, distribute or sell copies of this thesis in microform, paper or electronic formats.

L'auteur a accordé une licence non exclusive permettant à la Bibliothèque nationale du Canada de reproduire, prêter, distribuer ou vendre des copies de cette thèse sous la forme de microfiche/fiùn, de reproduction sur papier ou sur format électronique.

The author retains ownership of the copyright in this thesis. Neither the thesis nor substantial extracts fi-om it may be printed or otherwise reproduced without the author's pemiission.

L'auteur conserve la propriété du droit d'auteur qui protège cette thèse. Ni la thèse ni des extraits substantiels de celle-ci ne doivent être imprimés ou autrement reproduits sans son autorisation,

THECONTROL OF LOWER LIMB SENSORY INPUTS ASSOCIATED WTH A SEATED INVERTED-PENDULUM BALANCING TASK

Darcy Clair Bishop University of Guelph, 1999

Advisor: Professor WE. McIlroy

This thesis is an investigation of the control of sensory inputs during the performance of inverted-penddum balancing tasks. niree experiments were conducted to examine modulation of proprioceptive inputs during balance tasks of varying difliculty. It was hypothesized that (1) proprioceptive inputs to the cortex would be facilitated during a chailenged balance task, (2) such facilitation would not be observed at the spinal level, (3) cortical facilitation would be unaffected by peripheral inhibitory innuences, and (4) the availability of visual inputs would attenuate the facilitation of proprioceptive inputs.

The hypotheses were supported by the data. It is concluded that there is task-specific reweighting of sensory inputs to the coaex during chailenged balance control. This reweightuig may depend on the availability ofalternate sensory inputs. The increased

transmission of proprioceptive inputs to the cortex and the decreased transmission at the spinal level suggest that the cortex plays an important role in challenged balance, whereas

the role for the spinal (H) reflex appears to be less important.

Foremost, I would like to acknowledgethe efforts of my advisor Bill M c b y and extend many thanks for the guidance he provided tbughout the development and writing of this thesis. He was always wiIlIng to answer questions and ensured I stayed

focused on the questions 1was addressing with my research. Without his knowledge and technicd expertise it wotxid have ken much more diffïcult to accomplish my xsearch goals. 1 would also Lice to acknowledge the contniutions ofmy advisory cornmittee. To

John Brooke,I wouid like to extend thanks for sharing some of his wisdom with me. His lmowledge of general science and the arts was a resource 1feel privileged to have been able to tap ittto. The comments and contributions of Gary P d o w over the course of this thesis were also greatly appreciated The contributions of m y colieagues can not be overstated. Whether 1needed help with an experiment or just to tallc about research or Me, Rob and Aimee were always

there for me. For that 1wiu aiways be grateful. Thadcs guys, 1wish you aii the best in your fbture pursuits.

Last, but defïnitely not least, 1would Like to thank Lindsay. Her undying support and

understanding were pillars of strength for me during many stressfbl times. This was

especialiy true in the final stages ofpreparing this manuscript Thanks for k i n g there Linds.

Chapter FOE 4.0 Discussion...................................................................... 4.1 Balancing task effects......................................................... 4.2 Separation ofcontrolofcorticai and spinal afférent inputs............. 4.3 Effêcts ofperipheral conditionhg........................................... 4.4.Effects ofadditionai sensory input.......................................... 4.5 Geneialisabilityofthe results ............................................... . . 4.6 Limitanom ..................................................................... 4.7 Loci of&irent inflow moduIation ......................................... 4.8Conclusionsand hctional implications...................................

References.............................................................................

List of Tables 1.

Mean latencies (with standard deviatiom) ofthe SEP compnents (P 1, NI, P2,N2)dong with those for M waves and H rdexes.........

Page

26

List of Figaies

Page

of inputs h mthe muscle spindle (Ia) affierents Pathways for the 15 to the cortical and spinal Ievels.. .................................................. Experimental set-up.. ................................................................

17

Mean (d tibid ) nerve SEP (A) and soleus M wave and H reflex traces (B) fiom one subject for NO BALANCE,BALANCE and TJ3REATENED BALANCE conditions.. .............................................................

27

Mean amplitudes for Pl-N1 (A) aud P2-N2 (B) amplitudes @=IO) for the three task conditions h m experiment one.. ......................................

29

Mean M wave and H reflex amplitudes (with standard errors) for the three task conditions fiom experiment one @=IO). .............................

30

Mean PI-Nl (A) and P2-N2 (B) amplitudes (n=l) for the four task conditions fiom experiment two...................................................

33

Group M wave and H reflex amplitudes (n=7) for the four task - mean . conditions in experiment two.. ....................................................

34

Mean values of the grouped data (n=7) for Pl-NI (A) and P2-N2 (B) amplitudes for the four task conditions fiom experiment three...............

37

Mean values for Pl-N1 amplitudes nom the two subsets of subjects fiom expriment three.. ...................................................................

38

10. Group mean M wave and H reflex amplitudes (n=7) for the four task conditions fiom experiment three.. ...............................................

39

List of Abbmhtions

COM - center of mass BOS -base ofsupport CNS - central nervous system SOT - sensory 0rgani;rationtest SEP - somatosensoryevoked potential H reflex -H o f i a n reflex EEG -electroencephaIography EMG - electromyography ANOVA -analysis ofvariance SE - standard enor SD - standard deviation CI - confidence mtervai TA - tibialis antenor SOL - soleus POT - potentiometer TMS -trans-cranial magnetic Stimdation

Chapter One 1.0 Introduction:

The ability to maintainan upright posture is of paramount importance for the normal fllnctioning of e v e y h y We. It is the goal of balance control to maintainthis upright posture whether we are sitting, standing, walkhg or nmning. More formdy stated, the role of balance controi is to keep the body's center of mass (COM)over its base of support (BOS). Ifthe COM passes beyond the limits ofthe BOS the system will become

d will occur. Since falIs are a great risk factor for unstable and if this is not corrected a f injuries, especidy among the elderly, it is important to acquire an understanding of the mechanisms underlying how we maintain balance.

understanding can then,

hopefully, be used to develop better diagnostic and treatment options for the elderly and otherwise diseased patients d e r i n g h m balance disorders.

How is such complex whole-body balance control achieved? In order to maintain upright balance the central nervous system (CNS) requires constant monitoring of the

stability of the COM over the BOS. Displacements of the COM relative to the BOS can arise fiom intemally generated movements or h

m e x t e d perturbations. The control of

balance during the pex5omiance of internally generated perturbations associated with intended movements such as stepping could be achieved using predictive or anticipatory control. However, the control of balance subsequent to unpredictable perturbation events, such as a slip, is achieved using reactive or compensatory control. Whiie sensory inputs do have a role in predictive control of balance there is far greater reliance placed

on sensory inputs during reactive baiance control. This is attributable to the

unpredictability of the occurrence and characteristics of unexpected perturbations

encountered in everyday We- As a d

t of the importance of sensory inputs in the

readve control of balance there is a criticai d e phyed by sensory inputs in the maintenance of upnght postureThe monitoring of the COM stability by the CNS during reactive balance control is

mediated by a complex array of sensory inputs inc1iIding proprioceptive, cutaneous, visual, vestibuiar and haptic. These inputs together provide the information necessary to

detect and appropriately correct for disturbancesto stabiiity under a variety of circumstances. The importance of these sensory inputs is highiighted by the instability

experienced by patients with various sensory deficits. For example, patients with peripheral neuropathies affecfing the lower limb are unable to effectively make use of the ankle musculature for the generation of compensatory responses to d - s d e perturbations and are thus forced to use hip and other muscles to compensate (Horak et al. 1997). Conversely, patients d e r i n g fiom vestibdar loss are generally unable to

make adequate use of the hip strategy to correct for Iarger perturbations (Horak et al. 1997; Nashner et al. 1989). The consequence of both of these types of sensory deficits is

decreased stability accompanied by an increased risk of falling. An understanding of the importance of sensory inputs in balance control leads to

several important questions. How are these inputs used to monitor COM stability? Do they al1 play an equal role or do w e tend to rely on certain inputs more than others? If the latter is the case, what are the mechanisms by which the CNS achieves task- or situation-

specific adaptations to the weighting of specific sensory modalities? Understandingthe underlyhg neurophysioIogic mechanisms controlling the affierent inflow for the purposes

of successfbiiy maintaining stability is one of the long-tenn objectives of the present

work More specincaiiy, this study wiii explore the capacity of the C N S to gate/modulate the inflow of specifïc sensory inputs under specific balancing-task conditions.

An important preceding question is whether such moduiation of sensory inflow is

likely to occur during the control of balance in humans. In an effort to gain an understanding of how the various sensory inputs are organized, anumber of behavioural

experiments ushgvarious paradigms have examined the effects of variations in sensory inputs on balance conml (Nachner 1982; McCollum et 01- 1996). One of these paradigms, the Sensory ûrganïzation Test (SOT), was developed to examine changes in center of pressure sway under aitered visual and somatosensory conditions. These changes in sway patterns are used to make Inferences regarding the underlying CNS

regdation of balance control. The sensory conditions studied in the SOT included various combinations of visuai and somatosensory inputs including normal vision, normal somatosensory input, no vision, sway-referenced vision and sway-referenced somatosensory inputs. The sway-referenced conditions were designed to provide erroneous sensory feedback with respect to the baiancing task. The results of these studies Ied to the conclusion that the CNS g e n e d y relies on specifïc sensory inputs

under different circumstances for orientation information (Nashner 1982). Specificaliy, it

was concluded that somatosensory inputs are the most important for the control of reactive balance (Shumway-Cook and Horak 1986). This is not to say that the other sensory inputs are not important,just that the system appears to weight somatosensory inputs more heaviiy than visual and vestibuiar when ail are bctioning normaliy.

However, it was aiso ernphasized that the CNS must be adaptable so that when the available sensory information h m a given sensory modality is reduced or inaccurate an

alternative source of sensory input may be used for controlhg stability (McCoUum et al. 1996; Shumway-Cook and Horak 1986). The somatosensory inputs that appear to be most important for the regdation of upright stance are those fiom the lower leg, specifIdy, the muscle spindle inputs h m the muscles acting mund the ankle. Evidence for this cornes h m several sources.

Using a balancing task where subjects were required to maintain the stability of an

inverted-penduium (refened to as an ccequivaIentbody") using o d y movements of the

ankle, Fifipatrick et al. (1994a) concluded that proprioceptive signais h m receptors in the leg muscles are sufncient to maintain a stable upright stance. Diener et al. (1984)

showed that ischemic block produced by inflation of a sphygmomanometer cuffabove the knee impaired postural stability, although a similar block applied above the ankle did not. In a study that looked at the generation of compensatory response to a slip during gait, Tang et al. (1998) concluded that the inputs fiom the lower limb were the most

important for generating compensatory responses. Aithough it is generally agreed that somatosensory inputs are most important in reactive balance control (Shumway-Cook and Horak 1986; Fitzpatrick et al. 1994a) it is

noteworthy that changes in visual and vestibular inputs c m influence the control of stationary standing. Visual inputs appear, for the most part, to serve as a complimentary source of sensory information to proprioception in balance control (Hay et al. 1996).

Vision has been shown to be at least as accurate as propnoceptive inputs in determining the position of the adde (McCloskey et al.1983). However, it is generally believed that

visual inputs do not play a signincant role in the early stages of compensatory response generation although it can senre this purpose in the absence of proprioceptive inputs. The

role of vision wodd seem to be greater during lower fkquency c o d o n s such as those changes that occur in the later stages of the compensatory response (Nashner et aï-1989). Conversely, veshiular inputs are generally thought to play a d e as a standard against which the somatosensoryand visual inputs are compared (Nashner et al. 1982).

Stimulation of the vestibuiar apparatus via galvanic stimulation can induce balance reactions (Fitzpatrick et al. 1994a), however these inputs are most sensitive to low fiequency movements of the COM and are not held to contribuîe àgdicantlyto the generation of the rapid compensatory response (Nashner et al. 1.989). Wbiie behavioural studies have revealed the importance of sensory inputs for balance control and have provided evidence for the organhtion and use of them, they have contributed relatively Little hsight with regards to the acnial nemphysiology underlying balance control. Of paibicuiar interest is the occurrence ofmodality-specific increase in sensory response sensitivity that appears to occur under different task conditions. Such changes in sensitivity c m be achieved by altered central sensory transmission (by varying pre- or post-synaptic excitation or inhibition) or, in the case of muscle spindies, by

aitering fusimotor dnve to change receptor sensitivity (Prochazka et al. 1988; Aniss et alIWO). The important question for this study is whether there are changes evident in central sensory transmission that mirror the apparent switches in sensory weighting that have been suggested by b e h a v i o d observations?

Evidence for increesed muscle spindle sensitivity during ciifficuit balancing tasks has

been reported in cats (Prochazka et al. 1988) as well as in humans (Aniss et al. 1990). Aniss et al. (1990) showed specincally that in standing or maintainhg a static posture spindie activity is low, whereas when subject to unexpected perturbations the spindle

activity is greatly increased Whether this incresse in spindle sensïtivity is accompanied

by changes in the transmission characteristics of the saisory pathways carrying the

infirmation provided by these spindles is cuuentiy not known. It was the goal of the present work to examine changes in the transmission characteristicsof the sensory pathways conveying proprioceptive information k m the lower leg associated with changïng balancing task conditions. The transmission characteristics ofsensory inputs fromthe lower leg have been weii

studied at the spinal and cortical levels during non-balancing tasks. Transmission at these levels has been studied using potentials evoked by low threshold electrical stimulation of the tibial nerve at the ankle, the sural nerve, common peroneal and the tibial nerve at the

popliteal fossa. Of specinc interest to the present study is the relative contribution made

by muscle spincile afEerents from the extensor muscles around the ankle (specifïcaüy, soleus), which can be activated by stimulation of the tibial nerve at the level of the popliteal fossa. Stimulation of this nerve at low voltages 6 t s primarily the Ia afferents that carry the information fiom the muscle spindles in triceps surae (Halonen et al. 1989). Bio-potentials that arise fiom this stimulation are the spinai Hoilinan (H)

reflex and the cortical somatosensory evoked potential (SEP). Monitoring of these potentials allows for the observation of changes in the transmission characteristics of the pathways carrying the inputs while bypassing the receptor (muscle spindie) itself. The H

reflex represents motor activity in the muscle resulting fiom the monosynaptic

transmission fkom the Ia afTerents to the motor neurons of the same muscle. Transmission to the cortex is represented by the SEP. The same Ia afEerents that give rise to the H reflex send collaterai projections, which synapse on Clarke's column cells in the

lumbar spinal c o d The axons of the Clarke's coIumn neurons travel in the dorsai spinocerebellartractof the spinal cord and terminate at Nucleus Z in the braiastem. The Nucleus Z neurons then cross the midline in the medial Iemniscus and travel to the ventro-posterolaterai (VPL) nucleus of the Thalamus. The noal projection of tibial nerve inputs is then to the cortical pyramidal ceiis of Brodman area 3a The initial component

of the cortical SEP (PLNI)is held to represent the depolarkation of tiiese cortical

pyramidal c e k arïsing h m the direct arrivai ofthe & i t input at the somatosensory cortex (Gloor 1985). Modulation of the central sewry representation c m occur at ail levels of the neuraxis, fiom the spinal segmental refiex to the primary cortical receptive zones (MacKay and Crammond 1989). Modulatory Uinuences on semry transmission are generally thought to arise fiom two sources. The nrst type of modulation, refmed to as centriîugai gating, arises nom central descending idluences. Centrifùgal gating of

sensory inputs has been revealed by measuring the changes of both spinai (H) reflex (Collins et al. 1993)and cortical (SEP)representation (Staines et al. 1997a)of Ia af5erent input prior to or during movement Two recent studies have provided evidence of taslc

related centrifiigal-induced facilitation of sensory inputs to the cortex during movement of the upper (Knecht et al. 1993)and lower Iùnbs (Staines et al. 1997a). The results reported by Knecht et al. (1993)are however difncuit to ioterpret, as the f d t a t i o n they reported was the appearance of a new potential, during exploratory finger movements, that was not seen during normal movement However, the finciings of Staines et al.

(1997a) clearly demonstrate facilitation of the primary cortical SEP component (PLNI) prior to and during the performance of a foot-tracking task. In this study subjects were

required to track the movement of a foot k i n g moved passively with active movement of the other foot The authors reported fhdtation of the P LN1 component both before and

duxing movement and concluded that centrifügal factors are important in moduiating SEP gain required by the kinaesthetic demands of the task.

The second source is peripheral or centripetai modulation of Serent inputs. This type of modulation is induced by concurrent discharge o f a separate sensory input P e r i p h d y induced conditionhg of sensory inputs h m the lower Ieg has been well studied at the spinal level (Brwke et al. 1997)but only recently studied at the cortical level (Staines et al.199%).

One of the main sources of centrîpetal gating that is relevant

to balance control is that arising fiom muscle spindle discharge in the extemors of the

knee and hip consequent to passive and active movement (Brmke et al. 1997). During passive movement the observed gating is proposeci to arise £kompresynaptic inhibition of Ia afferent transmission (Brooke et al-1997). This mechanisrn is aiso felt to contribute at

ieast partly to the gating that occurs during active movement (Staines et al. 1997b), but

this is likely accompanied by centrifiigal influences (Capaday and Stein 1986). Evidence of task-specif7c gating of cortical representation of proprioception nom the lower limb, during non-balancing tasks?has led to the possibility that similar changes wül occur duriog balancing tasks. Chapin and Woodward (1982)have provided indirect

evidence for this in the cat. These authors reported that the activity in somatosensory cortical neurons was depressed durhg gait, but this depression was removed and the activity was increased during exploratory movements or locomotion over irregular

surfaces. Thus, it was suggested that sensory transmission might be increased in situations where it is required for adaptive responses to the environment (Prochazka

1989). It is p~sently proposed that the increased demand for proprioceptive cues to

maintain baiance when there is a threat of extemal perturbations will be paralieleci by an increase in the c o r t i d representation of these sensory inputs.

The objective of the present work is to examine the regdation of a f f i n t inputs during balance tasks of varying diff?culty. The general questions to be addresseci are as follows. (1) Does the cortex regulate the amount of sensory input it receives during the performance of baiancing tasks ofvarying difEcnlty? The proposition, as demonstrated

for non-balancing tasks, is that increased sensory demand would Iead to faciiitation of sensory inputs. (2) If the sensory inputs to the cortex are regulated differently with increasing task diffculty, how are these changes affected by peripheral inputs similar to those that arise fiom movement? Foliowing fiom the findings of Staines et al.(1997a),

which showed suppressionof movement induced depression of cortical afferent transmission, it is proposed that similar suppression of potentiai inhibitory influences would be evident during a challengeci balancing task. (3) Lastly, is this regulation stiIi evident when an additional relevant source of sensory information is available? Vision serves as a source of comphnentary seasory information when normal somatosensation is available. As such the reliance on proprioception is likely affiected by the presence or absence of vision. The ability to use an additionai (possibly redundant) source of sensory input wiil attenuate pathway gain as a result of reduced need for proprioceptive

information. The present study focused on two CNS loci, specincally, the primary sensory region of the cortex and the segmental spinal level. The H reflex represents transmission at the

spinal level and transmission a .the primary sensory regions of the cortex is represented

by the SEP. A seated inverted-penddum balancing task was used to explore the gating of A sensory inputs h m the Iower limb associated with increased balancing task dif£ïcuItyty

seated task was chosen in order to be able to remove vesti'bular influences h m the balancing task. This allowed us to focus on the isolated inputs fiom the lower limb and subsequentlyon the e f f ' of superimposing visuai inputs. The specific task requiEed subjects to maintain the position ofan inverted penduium with and without the threat of

extemal perturbations to the stabfity of thependulum. A similar task has ken used previously to study balance control mecbanisms (Fitzpatrick et al. 1994a).

Three experiments were conducted to address the following hypotheses.

Cortical transmission of afXerent input fiom the lower leg will be facilitated, as reflected by increased amplitude ofthe initiai SEP component, during a balancing task of increased diff?culty. Peripheral effects induced by concurrent stimulation o f h e e extensor

proprioceptors will not influence the facilitation of the SEPSobserved during the challenged balance task. The facilitation of cortical transmission of sensory inputs fiom the lower limb will

not be seen when additional relevant sensory input (vision) is avaiiable during the challenged balance task.

4) There wili be d i f f i t i a l modulation ofthe cortical and spioal transmissionof

afEerent input. This s e p t i o n will be chanicterùed by differentiai task-specific

modulationofH-reflex and SEP amplitudes to a common test stimulus.

Chapter Two

Methds 2.0 General Methods: 2.0.1 Subjects Seventeen volunteer subjects (nine women and eight men) ranging in age from 21 to 40 years participateci in one or more of the experiments. Each subject gave informed

consent and none reported any history of neummuscular deficits. The experimental procedures were approved by the University of Guelph Ethics Cornmittee for the use of

Human subjects. 2.0.2 Stimulation procedm

Somatosensory evoked potentids (SEPS) and soleus H reflexes were used to assess the transmission characteristics of the cortical and spinal afXerent pathways respectively.

These potentids were elicited by electrical stimulation of the tibial nerve at the popliteal fossa of the right leg in all experiments unless otherwise noted. A square wave pulse of

0.5 ms duration (Grass S88 with S N 5 stimulus isolation unit) with a 30 ms pre-stimulus delay was delivered over the tibial nerve via surface electrodes, with the anode d i a .

The impedance at the stimulating site was less than 10 kn,measured at 30 Hz (Grass

EZMS impedance meter). The intensity of the stimuius delivered to the nerve was rnaintained by monitoring soleus M wave magnitude at approximately 10%of the

maximal M wave (Mmax). The tibial nerve was selected for several reasons. Firstly, it contains the

proprioceptive inputs fiom ankle extemors, which are viewed as a prùnary sensory source in this task, and stimulation leads to responses arîsing from both spinal (Hreflex)

and cortical (SEP)levels (see Figure 1). Secondlysit is possible to bio-caliirate the

stimulus intensity by monitoring the magnitude of the M wave, which is held to represent the direct stimulation of motor efferents that are not susceptiible to potential modulatory effects (Brooke et al. 1997). Neurographic studies have shown that there is a strong

correlation between M wave magnitudes and low threshold afferent volleys (Abbruzzese et al, 1985; Fukoshima et al. 1982).

2.03 Recording procedures Electroencephalogmphic(EEG) scalp electrodes in an Electro-Cap system (ECI) were used to record SEPs. SEPs were recorded fiom Cz' (2 cm caudal to Cz) referenced to

Fpz' (2 cm caudal to Fpz) in accordance with the International 10-20 System for electrode placement (Jasper 1958). Cz' was chosen because it approximately overlies the primary somatosensory cortex for the lower limb (Homan 1987). Fpz' is typically

selected as the reference site for recording h m stimulation of peripheral nerves in the lower limb (Amer EEG Soc 1984). Furthemore, the first cortical component of the SEP due to tibia1 nerve stimulation was maximal with this placement of the active and reference electrodes. Electromyographic (EMG) recordings were obtained with surface Ag/AgCl electrodes oriented approximately 2 cm apart longitudindy over the predicted fiber paths of the soleus and tibialis anterior muscles. A common ground electrode was placed on the clavicle. Impedance at a i l EEG and EMG recording sites was less than 3kR and lOW, respectively, measured at 30 Efz (Grass EZM5 impedance meter). EEG and EMG recordings were ampmed (~40000,and ~ 1 0 0 0respectively) , and bandpass filtered (11O0 Hz and 3-300 Hi, respectively) with an isolated bioelectnc amplifier (SA

Instrumentation). The data were dipitized at an analog-digital interface (Keithley, DAS1800HC), with a sampling nite of 1000 Hz for a period of 150 ms,and stored on cornputer for subsequent analysis.

Midline

Nucleus Z Clarke's column

-

Cortex

A) SEP

B) H reflex

Figure 1: Pathways for the transmission of inputs nom the muscle spindle (Ia)

afEerents to cortical and spinal levels. (A) The somatosensory evoked potential (SEP) is recorded at the scalp and represents transmission at the cortical level. @) The H reflex is recorded at the muscle (soleus) and represents transmission fiom the la afFerent the alpha motor neuron (aMN)at the spinal level- These potentials both arise fiom the same

electrical stimulus applied at the popliteal fossa

2.0.4 General Set-ap

Subjects were seated in a slightly recliwd, rigid metal chair with the head and neck supported. Hip and knee angles were maintained at approximately 105' and 90° respectively. The subject's feet were spaced approximately 6-8 inches apart and secured with velcro straps to an inverted pendulum fmt-platfonn that couid rotate about the

plantar-, dorsi-flexion axis of the ankle (see Figure 1). The pendulum was approximately lm high and was weighted at the top with 5-8 kg depending on the subject Weights were also added to the back of the platform in order to reduce tibialis anterior (TA) activation, load soleus (SOL)slightly, and make the platfonn relatively easy to maintain in the upnght position.

The pendulum had a range of motion of approximately 15degrees of plantar-flexion

and 22 degrees of dorsi-flexion. The perturbations (plantar-flexion and dorsi-flexion) were induced manually by the experimenter and were delivered in a raudom order with

respect to timing, direction and magnitude. In order to reduce the potential effects of changing ankle position on the SEPSand H reflexes a potentiometer (POT)mounted on the axis of the ankle was monitored on an oscilioscope to ensure that stimuliwere

delivered at a constant ankle position throughout the experimental conditions.

Figure 2: The experimental set-up.

indicaîes the site of stimulation of the

tibia1 nerve in the popliteai fossa. EMG was recorded fiom soleus (SOL)and tibialis anterior (TA). SEPS were recorded from the scalp (Cz'). Subjects wore opaque goggles to eliminate visuai information in Experiments One and Two. The dashed iines represent

barriers that were put in place for Experiment Three to shieId the experimenter h m the subject during the eyes open conditions.

2.0.5 Data milysh

Mean traces fkom which SEP and EMG amplitudes were measured were comprised of 50-80 individual sampIes, v i d y inspecteci to be artifact k e . Artifàcts which

comtïtuted basis for trial rejection were; 1) the presence of large wave (alpha) EEG activity, 2) the presence of EMG noise c o n ~ t i n the g -EEGtraces, or 3) the presence of 60 IIz noise, arising fkom electrical sources. Ushg the average data, SEP amplitude measurements were taken fiom the peak of the first positive deflection (Pl) to the peak of the subsequent adjacent negative defiection (Nl), and fiom the peak to peak of the next

adjacent positive and negative potentiais (P2-N2), for d three experiments (see Figure 3A).

Noise bands and confidence interva2sfor SEPS

Pilot work by Staines (1997) to determine the number of samples required in an average for a reliable P LN1 SEP recording found the 95% confidence interval (CI) for Pl-Nl mean amplitudes (based on averages of 40 and 80 samples) to be 11.1% and 5.5%.

The noise band for EEG recordings in the present study with no stimulation and no movement was found to be approximately f 0.04 pV. With no stimuiation and vibration

the noise band increased to approximatelyt 0.05 pV.

M and H waves were analyzed by using the absolute peak to peak values fiom the average traces (see Figure 3B). Subsequent to analysis the values were normaiized as a percent of the maximal M wave (%Mm=) for graphitai representation. Pre-stimulus EMG anaiysis used the average ampiitude of full wave rectified EMG fiom 25 ms of the pre-stimulus interval.

Specific aprion cornparisons on a repeated measure d y s i s of variance (ANOVA), blocked on subjects, were used to test the research hypotheses for aii experiments. These cornparisons were made using the error mot mean square fiom the ANOVA, adjusted for the appropriate degrees of &dom (Kuehl1994). Scheffe's aposteriori test for multiple

cornparisons was used to test for any M e r differences between means. The signincance level was taken to be p i0.05. When necessary, Iogarithmic transfomiations of the data were used to ensure homogeneity of variance and normaiity of the residual ermrs to meet the assumptions required for the ANOVA.

2.1 Experiment One Balanchtg Tesk EHecB -direrence due fo the chdenge to mainfain stabiii#y

2.1.1 Subjects Seven subjects (three male and four femaie), ranging in age fiom 22 to 26 years (mean 23 -4 f 1.6) participated in the snidy. 2.1.2 Task conditions

This study involved three seated-bdmcing task conditions of varying difficuity. In this experirnent subjects wore opaque goggles to elimiaate visual inputs. The nrst task

required the subject to sit quietly with the pendulum locked in position (NO BALANCE).

In the second condition, the penddum was fiee to rotate but the subject was required to maintain the pendulum in a stable, upright position (BALANCE). In the last and most difficult condition, the subject was again ~quiredto maintain the penddum in an upright position but was instnicted to detect and correct for perturbations to the pendulum

(THREATENED BALANCE), The conditions were randomized and cokcted over six rounds of ten trials each for a total of sixty triais for each task condition.

2.13 Stimulation pmceànres In the THREATENED BALANCE condition the electrical stimuli (to elicit SEPSand

H reflexes) were delivered in the period between perturbations, when the subject had retumed the penduIum to the upright position and there was very little or no movement of the penduiumtlIll This was done to d u c e the potentid for codiounding effeçts of peripherai

af3erent discharge caused by the perturbations. This was determined by selectùig a time at which the angular position of the pendulum was stable. Ahhough the perturbation of

the penduium ais0 evoked cortical potentials that codd be recorded at the scalp these

were not used in the present experiment.

2.1.4 Data anaiysis Specinc apriori cornparisons (one-way contrasts on a repeated measure ANOVA) were used to test the research hypotheses.

202 Experirnent Two Influence of per@heraf condilioning on balanchtg task effects

2.2.1 Subjects Seven subjects (five male and two femaie), ranging in age fkom 21 to 40 years (mean 26.1 f 6.4) participated in this study.

22.2 Stimulation and recording procedures

In this experiment the electrical stimulus was delivered to the tibiai nerve at the popliteal fossa of the lefi leg as opposed to the right leg. This was due to constraints of

the physical set-

of the experimental apparatus. AU recordhg procedures were the

same as previously outlined

2.23 Task conditions Opaque goggles eliminated visual input aud extraneous noise was minimizad. In this experïment subjects were required to perfionn the BALANCE and THREATENED BALANCE tasks (outlined in Experiment One) with concurrent vibration of the vastus lateralis muscle. These two task conditions were couected in 4 or 5 interdigitated rounds of 15-20 trials each for a total of 75-80 triais per condition for d y s i s . The subjects aiso

performed BALANCE tasks,without vibration, before and &er the vibration trials.

These non-vibrated BALANCE tasks were coiiected in blocks of 60-80 trials separated f?om the vibration block by a ten-minute rest period. 2.2.4 Vibration

Vibration of the vastus lateralis muscle (to elicit quadriceps muscle spindle discharge) was applied manually by a second experimenter using a hand held vibrator (Poilenex

Power Massager). The vibration was applied continuously throughout the vibrated trials with the exception of several one-minute breaks given to the subject in order to avoid

fatigue effects. The vibration kquency was approximately 100 Hz with amplitude of 12 mm. These have been shown to be effective vibration characteristics for the

recruitment of muscle spinciles (Rolland Vedel 1982). 2.2.5 Data anaiysis

Specinc apriori cornpiirisons (one-way contrasts on a repeated mcasure ANOVA)

were used to test the research hypotheses.

2.3 Experiment Three

In/luence of visuol hpuf on balunchtg task enects

23.1 Subjeeb Seven subjects (four male and three fernale), ranghg in age fiom 21 to 27 years (mean 23.4

+2.5) participated in the study.

2.3.2 Task conditions In order to test the effécts of vision on task effects from Experbent One the same conditions were canied out with and without vision of the pendulum. In the vision conditions two separate shields were put in place so the subject couid not see the experimenter deliver the perturbations (see Figure l), but could see the penddum move. There were two subsets of subjects for this experiment These two subsets of subjects performed slightly dif5erent task conditions. The first group (n=3) performed the three tasks fiom Experiment One (NO BALANCE, BALANCE, and TKREAT'ENED

BALANCE) with eyes open and with eyes closed. The second set ofsubjects (n=4) only performed the BALANCE and THRJ2ATENED BALANCE tasks with eyes open and with eyes closed However, in order to address separate research questions, the second set of subjects was tested on the suml nerve as well as on the tibia1 nerve at the popliteal

fossa Thus, the first group of subjects had six separate experimental conditions, whereas the second group of subjects had eight experimental conditions. Note that the m a l nerve data is not reported in this thesis.

.

The task conditions were blocked into groups of 15-20 trials. These blocks were interdigitated, randomized and repeated four times for a total of 60-80 trials per condition

for anaiysis.

2.33 Data anaiysis Specinc aprion' wmparisons (one-way contrasts on a repeated measure ANOVA) were used to test the research hypotheses.

Chapter Tbm Resdts

3.1 Experiment One Balanchtg fusk effeets The data presented here represent pooled data including the subjects fiom Experiment

One (n=7), and subjects who performed the same task conditions in two fürther experiments ( ~ 3two , from Expe-ent

Three and one pilot subject) giwig a total of 10

subjects for the analysis. These data were pooled in order increase the sarnple size for

task comp~sons.This was possible since there was a subset of subjects in Experiment Three who pedonned the three task conditions fiom Experiment One. The average latencies across subjects for the eady SEP components fiom tibiai nerve stimulation are reported in Table 1for each experimental condition, together with the overall averages for

each SEP component These latencies dong with those for the M waves and H refîexes did not differ due to experimental conditions in this or the other two experiments.

Figure 2 shows the raw traces of 60 averaged (A) tibial nerve SEPS and (B) H

reflexes for the three task conditions (NO BALANCE, BALANCE, THREATENED BALANCE) fiom one subject. There was a facikition of the PLNI compnent in the

THREATENED BALANCE condition relative to the BALANCE condition. This facilitation is represented in the group means, (shown in Figure 4A with standard e m ) as a significant 33% increase in amplitude (F=5.0 p=0.0 19). Increases were seen in

seven out of ten subjects.

There was a 27% depression in the group mean Pl-N1 amplitude h mthe NO BALANCE task to the BALANCE tasIc, although this was not statistically significant

(F=2.0 p=0.09). This depression was aIso seen in seven ofthe ten subjects.

Table 1: Mean latencies (with standard deviations) of the SEP components (Pl, NI, P2, N2) dong with those for M waves and H reflexes. These values are within the normal reported ranges and did not vary signincantiy across task conditions or experiments.

LATENCES

NO BALANCE

BALANCE

THREAT BALANCE

Stimulus

NO BALANCE

BALANCE

THREATENED BALANCE

Stimulus

10 ms

NO BALANCE

BALANCE THREATENED -/* BALANCE

Figure 3: Mean (n=60) (A) tibial nerve SEP and (B) soleus M wave and H reflex traces fÏom one subject for each experimental condition (NO BALANCE,BALANCE,

THREATENED BALANCE). (A) SEP amplitude measurements for Pl-N1 were taken fiom the peak of the f h tpositive deflection (P 1) to the peak ofthe fïrst negative potentiai (NI). P2-N2 amplitudes were measured nom the second positive (P2) to the

subsequent negative (N2) peaks. The arrows indicate stimulus onset.

The amplitude of the secondary (Pî-N2) SEP cornpiex was slightiy Iower in the two balancing conditions relative to the NO BALANCE CONDITION (although not statisticaiiy signincant), but t h æ was no cliflierence in P2-N2 amplitude between the

BALANCE and THREATENED BALANCE conditions. Vigure 4B)Figure 5 shows that the H reflex amplitude was significantly depressed in the

THREATENED BALANCE condition relative to the NO BALANCE (1 1Yo) and the BALANCE (8%) conditions (F=1095fl.004, and F= 5.14 @.O36

respectively). The

amplitude of the H reflex was not sipifïcantiy different between the NO BALANCE and

BALANCE conditions (F=1.08 p 4 . 3 1). The H reflex data was log trdorrned to attain nomiality. The changes observed in the SEPSand H rdexes were iikely not amibutable to

changes in stimulus intensity, muscle contraction or ankle angle. The intensity of tibial nerve stimuli was maintained to eiicit a mean soieus M wave magnitude of 9.8

+ 5.2%

Mmax and was not significautly dif5erent between any of the task conditions (ail p0.05). Pre-stimulus soleus EMG activity was maintahed at approxhmtely 25.OpV f 1.3 pV and was not significantly different between any of the task conditions. The position of

the ankle (as measured by a potentiometer) was constant across the BALANCE and

THEEATENED BALANCE conditions. The average difference between the two conditions was only 0.16' and was not statisticdy signincaot. However, the position of the adde was signincantly different in the NO BALANCE condition compared to the other two, although there was a 0.93' merence with BALANCE and a 1.09O ciifference with THREATENED BALANCE.

e

nobalance

balance

nobalance

balance

th matbalance

th reatbalance

Figure 4: Mean amplitudes (with standard enors) of (A) Pl-Nl and (B) P2-N2 for 10 subjects across the three experimental conditions. There was a sipnincant increase in Pl-

N1 amplitude h m the BALANCE to the THREATENED BALANCE task wO.05). P2-N2 amplitudes did not change significantly.

+ denotes significance at

0.10.

* denotes signifïcance at p