Exp Brain Res DOI 10.1007/s00221-008-1463-0
R ES EA R C H N O T E
Galvanic vestibular stimulation speeds visual memory recall David Wilkinson · Sophie Nicholls · Charlotte Pattenden · Patrick KilduV · William Milberg
Received: 29 April 2008 / Accepted: 5 June 2008 © Springer-Verlag 2008
Abstract The experiments of Alessandro Volta were amongst the Wrst to indicate that visuo-spatial function can be altered by stimulating the vestibular nerves with galvanic current. Until recently, the beneWcial eVects of the procedure were masked by the high levels of electrical current applied, which induced nystagmus-related gaze deviation and spatial disorientation. However, several neuropsychological studies have shown that much weaker, imperceptible currents that do not elicit unpleasant sideeVects can help overcome visual loss after stroke. Here, we show that visual processing in neurologically healthy individuals can also beneWt from galvanic vestibular stimulation. Participants Wrst learnt the names of eight unfamiliar faces and then after a short delay, answered questions from memory about how pairs of these faces diVered. Mean correct reaction times were signiWcantly shorter when sub-sensory, noise-enhanced anodal stimulation was administered to the left mastoid, compared to when no stimulation was administered at all. This advantage occurred with no loss in response accuracy, and raises the possibility that the procedure may constitute a more general form of cognitive enhancement. Keywords
Vision · Memory · Sensory stimulation
D. Wilkinson (&) · S. Nicholls · C. Pattenden Department of Psychology, University of Kent, Canterbury, Kent CT2 7NP, UK e-mail:
[email protected] P. KilduV · W. Milberg New England GRECC, VA Boston Healthcare System, Boston, MA, USA W. Milberg Department of Psychiatry, Harvard Medical School, Boston, MA, USA
Introduction The speed at which a visual memory can be recalled is aVected by a host of factors, including the manner of acquisition and storage, as well as the more transient inXuences of mood and pharmacological agent. However, spontaneous increases in speed are diYcult to achieve without either a loss in accuracy or greater conscious eVort. Here we show that low levels of galvanic vestibular stimulation (GVS), a simple procedure that delivers transcutaneous current to the vestibular nerves via electrodes placed on the part of the scalp that overlies the mastoid bones (Coats 1972), can speed memory recollection without such compromise. The vestibular organs of the inner ear convey information about the angular and linear acceleration of the head. This information is used to help maintain posture and ocular-motor reXex control, and also helps generate egocentric representations of space that are important for perception and memory (see Fitzpatrick and Day 2004). The eVects of vestibular processing on memory are compellingly demonstrated by the reduced spatial memory recall of astronauts under conditions of microgravity (Watt 1997), as well by the more commonly observed eVects of vestibular disease. In one recent study, 85% of patients with damage to the peripheral vestibular organs reported memory loss of some sort in the absence of more general intellectual impairment (Grimm et al. 1989). Such patients show impaired performance on virtual versions of the Morris water maze task (Brandt et al. 2005), and are troubled when they must make memory guided saccades (Tian et al. 2000) or recognise objects from memory (Zheng et al. 2004). Other tasks that involve spatial strategies such as counting backwards, visual imagery and mental arithmetic can also be aVected (see Hanes and McCollum 2006). These behavioural eVects are thought to arise from the dense connections between
123
Exp Brain Res
vestibular brainstem nuclei and regions in both medial temporal lobe, that contain structures necessary for spatial and non-spatial memory (Broadbent et al. 2004), and lateral temporal–inferior parietal lobe that are important for multisensory integration (Brandt and Dieterich 1999). In view of this background literature, it is perhaps surprising that few have attempted to improve memory performance by externally stimulating the vestibular nerves. As far as we are aware, only one group of researchers have attempted such a feat. In this experiment, Bächtold et al. (2001) injected cold water into the external ear canals of healthy participants who at the same time were asked to memorize the speciWc locations of diVerent objects. This procedure, known as caloric vestibular stimulation, releases convection currents that activate the balance organs in a manner similar to natural movement (for a comprehensive review of the eVects of caloric vestibular on cognition see Miller and Ngo 2007). Intriguingly, participants were faster at correctly recalling the location of objects during right caloric stimulation than either left stimulation or no stimulation at all. Regrettably, the procedure is very impractical, eliciting the severe side-eVects of sickness, dizziness, disorientation, and horizontal nystagmus. The aim of the present study was to establish whether improvement could be found using a much more tolerable, and somewhat simpler, means of vestibular stimulation— galvanic vestibular stimulation (GVS). GVS has mostly been used to probe the role of the vestibular system in autonomic control and only recently has its eVect on volitional processes been explored (Fitzpatrick and Day 2004). The procedure activates the vestibular nerves via small electrical currents that are applied to the overlying scalp, and elicits similar changes to regional cerebral blood Xow as its caloric counterpart (Bense et al. 2001). Previous GVS studies have, however, applied relatively high, super-sensory currents that invoke similar side eVects to caloric stimulation, as well as cutaneous discomfort at the sites of stimulation (e.g. Saj et al. 2006; Lenggenhager et al. 2007). This has led to the idea that while GVS might provide an additional means of examining the function of the vestibular system, it oVers little in the way of cognitive enhancement. Recent neuropsychological studies have, however, shown that this assumption does not hold when the current is reduced to a level that is too low to elicit side-eVects. At these subthreshold levels, spontaneous improvements in the symptoms of prosopagnosia (Wilkinson et al. 2005), hemi-spatial neglect (Rorsman et al. 1999) and autonomic dysfunction (Yamamoto et al. 2005) have been observed. Although these facilitatory eVects are not well understood, they show that GVS aVords restorative qualities that have hitherto been masked by the super-threshold currents applied. Recent developments in biomedical engineering indicate, paradoxically, that the signal-to-noise ratio of sub-sensory
123
currents can be increased by adding a small, non-zero level of background noise to the waveform (e.g. random electrical interference) (Collins et al. 2003). The degree of random variance is determined by specifying both the mean and standard deviation (where 0 = no noise) of the waveform, and follows a Gaussian distribution. At an intermediate level of variability, the output of sensory neurons becomes stochastic which has been shown to speed signal transmission and slow habituation (see Moss et al. 2004). Stochastic Resonance (SR) was originally proposed in the context of global climate modelling as a possible explanation for the periodic recurrences of the earth’s ice-ages (Nicolis 1982). However, the non-linear threshold detection mechanism that neurons employ, along with the noisy environment in which they reside, led to the idea that SR may also play a role in neural signal detection. The idea is that low-level electrical noise causes changes in receptor membrane potentials, making neurons more likely to discharge an action potential in the presence of a weak signal (see Moss et al. 2004), and, given the unpredictable nature of the signal, less likely to induce learning and anticipatory eVects in the subject (Balter et al. 2004). In line with SR theory, a number of studies have shown that small levels of input noise can sharpen the responsiveness of peripheral, spinal and cortical neurons. At the psychophysical level, Gaussian noise lowers the absolute threshold for tone detection in normal hearing individuals (Zeng et al. 2000), enhances the detection of weak tactile stimuli (Collins et al. 1997), and increases contrast sensitivity and luminance detection (Piana et al. 2000; Riani and Simonotto 1994). In the present context, we therefore reasoned that any beneWcial eVects of sub-sensory GVS on normal cognition would most likely be evident when the signal was convolved with noise. In the following sections, we Wrst report an experiment (Experiment 1a) in which 12 healthy participants performed a face imagery task during sub-sensory, noiseenhanced GVS. To demonstrate whether the observed eVects depended on the inclusion of electrical noise, we then brieXy report the results of a control experiment (Experiment 1b) in which identical procedures were administered to a new group of 12 participants, but with the exception that a non-noisy signal (i.e. electrical current of constant amplitude) was discharged. The behavioural task resembled that conducted by Barton and Cherkesova (2003) in which participants Wrst learnt the names of eight unfamiliar faces, and then after a rest interval, answered questions from memory about how pairs of these faces diVered at either the conWgural (e.g. ‘Does John have a fuller face than Barry?’) or featural (e.g. ‘Does John have thinner lips than Barry?’) level. Our decision to administer a face processing task, as opposed to an object–location association task in which caloric stimulation has been shown to boost performance (Bächtold et al. 2001), was based on
Exp Brain Res
our earlier success in remediating prosopagnosia via GVS (Wilkinson et al. 2005). In the Barton and Cherkesova study, patients with left hemisphere lesions were most impaired at answering questions about local features while patients with right hemisphere lesions were most impaired at questions about global/conWgural properties. The lateralisation of local and global processing Wnds much support in the psychological literature (see Rakover 2002), and given that hemispheric activation is greatest on the side nearest the cathode (Dieterich et al. 2003; Fink et al. 2003), we therefore counterbalanced the side of the head on which the cathode and anode were placed. We hoped (over-optimistically as it turned-out) that this procedure might unveil an interaction between anodal placement and speed of conWgural/featural recall. Of more importance, of course, was whether either stimulation condition (anode left-cathode right; anode right-cathode left) generated shorter and/or more accurate responses than the sham in which participants were falsely informed that they were receiving sub-sensory stimulation.
Method Participants Twenty-four volunteers were recruited from the University of Kent Psychology department participant pool. The Wrst 12 participated in Experiment 1a (noise-enhanced GVS), and the second 12 participated in Experiment 1b (constant GVS). All had normal or corrected-to-normal vision, and reported no neurological, vestibular or hearing impairment. The study was approved by the ethics panel of the University of Kent Psychology department, and all participants gave their written informed consent prior to study. All participants were right-handed, as assessed by the Briggs and Nebes (1975) Handedness Questionnaire. Stimuli Eight diVerent, unfamiliar, male faces were drawn from the face database stored at the Max-Planck Institute for Cybernetics, Tuebingen, Germany. The faces appeared in greyscale, were clipped at the hairline, and measured approximately 6 cm by 5 cm. All eight faces were printed on a single piece of A4 paper, and were given a hypothetical name that was printed alongside. Procedure Learning phase Participants sat in a quiet room and were given as long as they needed to learn the names of the eight faces. Partici-
pants only proceeded to the recall phase once they had accurately identiWed all faces via confrontation naming, and were given further time to re-inspect the faces if a mistake was made. The time taken to learn the face-name pairings was not recorded. Recall phase After a ten minute interval, participants answered questions from memory about how pairs of the faces diVered. The questions were adapted from Barton and Cherkesova (2003) to Wt the faces chosen from the stimulus database, 20 of which encouraged participants to focus on the featural properties of faces (e.g. ‘Does John have thinner lips than Barry?’), and 20 of which encouraged participants to focus on the conWgural properties (e.g. ‘Does John have a fuller face than Barry?’) (see Fig. 1a). The questions were presented auditorily and participants responded verbally via a ‘yes’ or ‘no’. The time between question oVset and response onset was recorded electronically, and the next question followed 2 s after response. Three blocks of trials were administered, each containing the same 40 questions in random order. Each block of trials was matched to a diVerent stimulation protocol; (1) anode left-cathode right, (2) anode right-cathode left, (3) sham. The order of stimulation was counterbalanced using Latin Squares, and participants were blinded to the stimulation protocol administered. Stimulation protocol Bipolar current (opposite current applied to each electrode) was delivered through a pair of 3 cm2 carbon-rubber, selfadhesive, disposable stimulating electrodes (ComfortEase, Empi Inc.) placed over the patient’s mastoid processes. To ensure complete electrical contact with the electrodes, surrounding skin was cleansed with an alcohol swab and conductive Tac-gelTM coated on the undersides of the electrodes. The electrodes were connected via a fast-acting fuse box (3 mA) to a stimulus isolation unit which was limited to 3 mA. Direct current was delivered to the isolation unit using National Instruments LabVIEW 6.0 and a dual output Microstar D/A board. In Experiment 1a, mean signal intensity was set at 90% of cutaneous threshold. Electrical noise signals had a frequency of 1,000 Hz, followed a Gaussian distribution, and were set at 0.25 of mean signal intensity (group mean intensity was 0.8 mA). Sensory threshold was determined using the stair-case method described in Wilkinson et al. (2005) and Rorsman et al. (1999). Electrical current was initiated at 0.1 mA with the degree of random variance set to 25% of the mean signal intensity. Mean signal intensity was then increased in 0.1 mA intervals for 10 s periods
123
Exp Brain Res
Results
a
Experiment 1a: noise-enhanced GVS
John
b
Barry 100
3
90 80 70
2
60 50
1.5
40 1
% errors
Mean correct reaction time (secs)
2.5
30 20
0.5
10 0
0 anode left
anode right
sham
Stimulation featural error
configural error
featural RT
configural RT
Fig. 1 a Example stimuli, b mean correct reaction time and error rates (with standard error bars) during noisy GVS and sham
A 2(Question type: featural vs. conWgural) £ 3(Stimulation: anode left vs. anode right vs. sham) repeated-measures ANOVA showed that mean correct vocal reaction times were signiWcantly shorter when the anode was placed over the left mastoid compared to the sham (F(2,22) = 4.2, P < 0.05, d = 0.3), with no loss in accuracy (F(2,22) = 0.8, P = 0.49, d = 0.06) (see Fig. 1b). This pattern was found in 11 of the 12 participants. The eVect was present for both featural and conWgural questions, though responses in the conWgural condition were shorter (F(1,11) = 4.7, P < 0.05, d = 0.3), and more accurate (F(1,11) = 9.3, P < 0.05, d = 0.5). Pair-wise comparisons ( = 0.05) showed that although anode right trials were not signiWcantly diVerent from anode left trials, they were also no diVerent from the sham. Experiment 1b: constant GVS 2(Question type: featural vs. conWgural) £ 3(Stimulation: anode left vs. anode right vs. sham) repeated-measures ANOVAs showed that both mean correct vocal reaction times (F(2,22) = 0.7, P = 0.68, d = 0.04) and response accuracy (F(2,22) = 1.3, P = 0.3, d = 0.1) were unaVected by the type of stimulation administered. As before, responses in the conWgural condition were shorter (F(1,11) = 15.9, P < 0.05, d = 0.61) and more accurate (F(1,11) = 32.7, P < 0.05, d = 0.75) than in the featural condition. The interaction between Stimulation and Question type failed to reach signiWcance in either the RT or response accuracy analyses (F < 1.0).
Discussion until the patient reported a mild tingling sensation at the electrode sites. To conWrm the precise level of threshold, the current was then be reduced by 0.3 mA and ‘staircased up’ once more to conWrm threshold. Threshold was determined at the start of each block. In the sham block, patients wore the electrodes and went through the same threshold procedure, after which they were falsely informed that they would receive sub-sensory levels of current during the subsequent block. In Experiment 1b, mean signal intensity was again set at 90% of cutaneous threshold but now with the noise signal set at 0 (e.g. zero). Sensory threshold was then determined in the same way as above. The mean sensory threshold of participants was found to be the same (0.8 mA), across the noise and no-noise experiments (t(22) = 0.7, P = 0.5, d = 0.04).
123
Compared to the sham condition, decisions were more quickly reported when anodal and cathodal stochastic currents were applied to the left and right vestibular nerves, respectively. Across the group as a whole, participants were approximately 0.5 s quicker to respond than in the sham condition. By contrast, non-noisy GVS did not speed performance. Previous studies have shown that sub-sensory GVS can improve acquired disorders of vision, but as far as we know, this is the Wrst to show a beneWcial cognitive eVect in neurologically healthy individuals. Why did low-level GVS shorten responses? At this stage we can only speculate, but one clue may lie in the fact that left anodal vestibular stimulation preferentially increases blood Xow to temporal and parietal structures in the right hemisphere (Dieterich et al. 2003; Fink et al. 2003). As
Exp Brain Res
mentioned, these structures are strongly associated with face processing (e.g. Wada and Yamamoto 2001) which would have been strongly engaged by a face imagery task. Earlier we predicted that the degree of improvement during left anodal stimulation would be greater for conWgural compared to featural questions. It is unclear why this diVerence did not emerge, but we note that other aspects of the task that were relevant to both kinds of question, such as the retrieval and comparison of relatively unfamiliar faces, rely on processes that are preferentially lateralised to right hemisphere (Laeng and Rouw 2001; Rossion et al. 2003). Likewise, many participants reported using the relative positions of each face on the test sheet to help remember which face paired with which name. Again, spatial strategies of this nature may be preferentially lateralised to right hemisphere (see Wilkinson and Donnelly 1999) so could also have been enhanced by left anodal stimulation.1 One other possibility is that the advantage reXected some kind of non-speciWc arousal. Given the sub-sensory nature of stimulation, we can dismiss motivatory and strategic factors that might have come into play had participants known when, and on what side, they were being stimulated. However, perhaps GVS ‘up-regulated’ activity across the brain and in fact led to widespread gain? Despite the appeal of such an account, we remain sceptical because if this were true then both anode right and anode left trials should have improved performance. Likewise, the remediation of particular neuropsychological deWcits seems to rely on diVerent anodal placements—the remediation of neglect relies on left anodal placement (Rorsman et al. 1999), while the remediation of prosopagnosia occurs only after both hemispheres have received anodal stimulation (Wilkinson et al. 2005). In fact, several other lines of evidence also give reason to doubt a non-speciWc eVect. First, GVS induces an asymmetry of the optokinetic after-nystagmus, which indicates an eVect on visual orientation that unspeciWc alerting stimulation does not (see Magnusson et al. 1985). Second, the beneWcial eVects of low-level GVS are task-speciWc; in neglect patients the procedure improves line crossing but not symbol cancellation (Rorsman et al. 1999). Third caloric vestibular stimulation, which exerts a similar eVect on central neural response, temporarily alleviates right hemispatial neglect and yet has no positive impact on the other left hemisphere disorder of aphasia (Vallar et al. 1995). In sum, the observed interactions between GVS, task and hemisphere suggest that the eVects are more circumscribed than would be predicted by a general arousal eVect.
1 We thank one of the reviewers for pointing out that right anodal stimulation might have been more eVective had it been applied during the learning phase in which left hemisphere verbal-semantic processes would have been needed to link names to faces.
A Wnal issue concerns the role played by electrical noise in bringing about the faster recall. Previous studies have shown that the stochastic resonance induced by rapid Xuctuations in signal amplitude invigorates the response of peripheral and central receptors to sub-sensory currents (Moss et al. 2004). In line with this, the reaction time advantage observed here was only present when the subsensory signal was convolved with noise. Importantly, the mean sensory threshold of participants was found to be the same (0.8 mA), across the noise and no-noise (e.g. constant GVS) experiments, implying that the mean signal intensity (90% of cutaneous threshold) across the two experiments was matched (though of course the amount of signal variation about that mean diVered). This result indicates that careful attention must be given to the basic signal characteristics employed; sub-sensory currents may prove ineVective if not convolved with noise while super-sensory currents (noisy or non-noisy) invoke discomfort and distraction so are similarly unsuitable. In sum, we have shown that face recall can be speeded if accompanied by low level electrical stimulation of the vestibular nerves. We note that improvement is contingent on both appropriate placement of the anodal and cathodal electrodes and the inclusion of electrical noise. Given that GVS is low cost, simple to apply, requires no active involvement of the participant, and can be easily packaged into a miniature device, we propose that further studies are warranted to both optimise the frequency and duration of stimulation, and to establish the range of cognitive behaviours that respond favourably to this long-established, albeit underapplied, technique. Acknowledgments We thank Jason Barton for providing the questions for the memory task, and Profs. Andrew Derrington and Howard Bowman for helpful comments on an earlier manuscript.
References Bächtold D, Baumann T, Sándor P, Kritos M, Brugger P (2001) Spatial- and verbal-memory improvement by cold-water caloric stimulation in healthy subjects. Exp Brain Res 136:128–132 Balter S, Stokroos R, Eterman R, Paredis S, Orbons J, Kingman H (2004) Habituation to galvanic vestibular stimulation. Acta Otolaryngol 124:1–5 Barton J, Cherkesova M (2003) Face imagery and its relation to perception and covert recognition in prosopagnosia. Neurology 61:220–225 Bense S, Thomas S, Yousry T, Brandt T, Dieterich M (2001) Multisensory cortical signal increases and decreases during vestibular galvanic stimulation (fMRI). J Neurophysiol 85:886–899 Brandt T, Dieterich M (1999) The vestibular cortex: its locations, functions, and disorders. Ann N Y Acad Sci 871:293–312 Brandt T, Schautzer F, Hamilton D, Brüning R, Markowitsch H, Kalla R, Darlington C, Smith P, Strupp M (2005) Vestibular loss causes hippocampal atrophy and impaired spatial memory in humans. Brain 128:2732–2741
123
Exp Brain Res Briggs G, Nebes R (1975) Patterns of hand preference in a student population. Cortex 11:230–238 Broadbent N, Squire L, Clark R (2004) Spatial memory, recognition memory and the hippocampus. Proc Natl Acad Sci 101:14515– 14520 Coats A (1972) The sinusoidal galvanic body-sway response. Acta Otolaryngol 74:155–162 Collins J, ImhoV T, Grigg P (1997) Noise-mediated enhancements and decrements in human tactile sensation. Phys Rev E 6:923–926 Collins J, Priplata A, Gravelle D, Niemi J, Harry J, Lipsitz L (2003) Noise-enhanced human sensorimotor function. IEEE Eng Med Biol Med 22:76–83 Dieterich M, Bense S, Lutz S, Drzezga A, Stephan T, Bartenstein P, Brandt T (2003) Dominance for vestibular cortical function in the non-dominant hemisphere. Cereb Cortex 13:994–1007 Fink G, Marshall J, Weiss P, Stephan T, Grefkes C, Shah N, Zilles K, Dieterich M (2003) Performing allocentric visuospatial judgments with induced distortion of the egocentric reference frame: an fMRI study with clinical implications. NeuroImage 20:1505–1517 Fitzpatrick R, Day B (2004) Probing the human vestibular system with galvanic vestibular stimulation. J Appl Physiol 96:2301–2316 Grimm R, Hemenway W, Lebray P, Black F (1989) The perilymph Wstula syndrome deWned in mild head trauma. Acta Otolaryngol 464:1–40 Hanes DA, McCollum G (2006) Cognitive–vestibular interactions: a review of patient diYculties and possible mechanisms. J Vestib Res 16:75–91 Laeng B, Rouw R (2001) Canonical views of faces and the cerebral hemispheres. Laterality 6(3):193–224 Lenggenhager B, Lopez C, Blanke O (2007) InXuence of galvanic vestibular stimulation on egocentric and object based mental transformations. Exp Brain Res 184:211–221 Magnusson M, Pyykkö I, Jäntti V (1985) EVect of alertness and visual attention on optokinetic nystagmus in humans. Am J Otolaryngol 6:419–425 Moss F, Ward L, Sannita W (2004) Stochastic resonance and sensory information processing: a tutorial and review of applications. Clin Neurophysiol 115:267–281 Miller SM, Ngo TT (2007) Studies of caloric vestibular stimulation: implication for the cognitive neurosciences, the clinical neurosciences and neurophilosophy. Acta Neuropsychol 19:183–203 Nicolis C (1982) Stochastic aspects of climatic transitions: response to a periodic forcing. Tellus 34:1–9
123
Piana M, Canfora M, Riani M (2000) Role of noise in image processing by the human perceptive system. Phys Rev E 62:1104–1109 Rakover S (2002) Featural vs. conWgurational information in faces: a conceptual and empirical analysis. Br J Psychol 93:1–30 Riani M, Simonotto E (1994) Stochastic resonance in the perceptual interpretation of ambiguous Wgures: a neural network model. Phys Rev Lett 72:3120–3123 Rossion B, Schiltz C, Crommelinck M (2003) The functionally deWned right occipital and fusiform “face areas” discriminate novel from visually familiar faces. Neuroimage 19:877–883 Rorsman I, Magnusson M, Johansson B (1999) Reduction of visuospatial neglect with vestibular galvanic stimulation. Scand J Rehabil Med 31:117–124 Saj A, Honore J, Rousseaux M (2006) Perception of the vertical in patients with right hemispheric lesion: eVect of galvanic vestibular stimulation. Neuropsychologia 44:1509–1512 Tian J, Cane B, Demer J (2000) Vestibular catch-up saccades in labyrinthine deWciency. Exp Brain Res 131:448–457 Vallar G, Papagno C, Rusconi M, Bisiach E (1995) Vestibular stimulation, spatial hemineglect and dysphasia, selective eVects. Cortex 31:589–593 Wada Y, Yamamoto T (2001) Selective impairment of facial recognition due to a haematoma restricted to the right fusiform and lateral occipital region. J Neurol Neurosurg Psychiatry 71:254–257 Watt D (1997) Pointing at memorized targets during prolonged microgravity. Aviat Space Environ Med 68:99–103 Wilkinson D, Donnelly N (1999) The role of stimulus factors in making categorical and coordinate spatial judgments. Brain Cogn 39:171–185 Wilkinson D, Ko P, KilduV, McGlinchey R, Milberg W (2005) Improvement of a face perception deWcit via subsensory galvanic vestibular stimulation. J Int Neuropsychol Soc 11:925–929 Yamamoto Y, Struzik Z, Soma R, Ohashi K, Kwak S (2005) Noisy vestibular stimulation improves autonomic and motor responsiveness in central neurodegenerative disorders. Ann Neurol 58:175– 181 Zeng F, Fu Q, Morse R (2000) Human hearing enhanced by noise. Brain Res 869:251–255 Zheng Y, Darlington C, Smith P (2004) Bilaterlal labyrinthectomy causes longterm deWcit in object recognition in rat. NeuroReport 15:1913–1916
