Visual-vestibular stimulation influences spatial and non-spatial cognitive processing

Abstract

This study investigated the impact of visual-vestibular stimulation on performance of an auditory information processing task in young and older adults. Performance on a spatial choice reaction time task was compared to performance on a nonspatial choice reaction time task. The tasks were performed during simultaneous rotational and moving visual stimulation. The non-spatial task was an auditory frequency discrimination task while the spatial task was a right-left lateralization task. Visual and vestibular conditions consisted of a non-movement baseline, sinusoidal earth-vertical axis rotation (EVAR) in darkness, off-vertical axis rotation (OVAR) in darkness at a constant velocity, OVAR in darkness with a sinusoidal profile, EVAR with a lighted visual surround, constant velocity optokinetic stimulation, and sinusoidal optokinetic stimulation. Baseline reaction times were subtracted from reaction times during each stimulus condition to yield “task cost”, which was analyzed statistically. Subjects were healthy young (n = 20; 24 ± 2.7 yrs; 10F) and older (n = 29; 73 ± 6.0 yrs; 18F) adults. Results indicated that task cost was affected by the visual-vestibular condition and a task × condition interaction. There was no main effect of task or age group and no significant interaction with age. Otolithic stimulation and visual stimulation were associated with greater task cost compared to semicircular canal stimulation. Combining semicircular canal with otolithic or visual stimulation had no additional effect beyond otolithic or visual stimulation alone. This pattern of task cost being larger for otolith or visual vs. semicircular canal stimulation was found for both the spatial and non-spatial tasks. The significant interaction between condition and task type revealed that the task cost for the spatial task was larger than the task cost for the non-spatial task during visual conditions but not during non-visual conditions although the visual and non-visual conditions were not entirely comparable. This study suggests that interference of vestibular stimulation with cognitive processing is especially prominent for otolithic and visual stimulation although the strength of the various visual-vestibular stimuli may not have been uniform. Also, spatial tasks are more affected than non-spatial tasks during visual stimulation and interference between vestibular stimulation and cognitive processing is not age dependent for these relatively easy tasks.

1. Introduction

Vestibular stimulation is known to adversely influence aspects of ongoing cognitive processing [5, 17]. This effect is similar to that seen in the postural control system when challenges to upright balance are combined with a cognitive task [12,13,16]. The basis for this interference between cognition and balance is uncertain but may relate to a bottleneck or limited capacity in the processing of competing attention-requiring tasks. Our previous and ongoing research has focused on the role of aging in information processing-balance/vestibular interference [5,12]. A previous study that combined visual-vestibular stimuli with auditory reaction time tasks showed that reaction time was significantly affected by vestibular stimulation and ocular pursuit in both older and young subjects [5]. Since the interference by vestibular stimulation with information processing was the same across three different reaction time tasks that differed in their central processing and motor characteristics, we surmised that interference occurred early in the processing stream as stimuli competed for a process with limited capacity, possibly in the temporal lobe. This notion is supported by recent PET studies showing cortical inhibitory sensory-sensory interaction and reduced excitability of V5/MT during vestibular stimulation [1,3, 14]. We also found that performing an auditory information processing task altered the timing, i.e. the phase, of the vestibulo-ocular reflex, as compared to that measured using a “standard” clinical alerting task [5]. A follow-on study indicated that the sensory vs. motor aspects of the cognitive task influenced this effect on the vestibulo-ocular reflex [15]. Our previous research has shown that the effect of age was small; age did not influence interference except during otolithic stimulation wherein older persons were influenced less than younger persons, presumably because of a decreased vestibular sensitivity with increased age. The motivations for the present study included a further exploration of the effect of age by including persons significantly older that the persons in our prior study, i.e., persons aged 75–85. Also, given the importance of the vestibular system in spatial orientation, and previous reports of interference between vestibular stimulation and cognitive tasks with a spatial component [7,8] we wished to explore the impact of the spatial properties of the information processing task on interference created by the vestibular stimulation in older subjects. For this study we increased the magnitude of the otolithic stimulation to account for the reduced sensitivity in older subjects found in our prior study.

2. Methods

2.1. Subjects

Data are reported from 29 older subjects (65–83 yrs, mean 73 ± 6.0 yrs, 18 F) and 20 young subjects aged (21–30 yrs, mean 24 ± 2.7 yrs, 10 F). Each subject provided informed consent approved by the Institutional Review Board of the University of Pittsburgh, whose guidelines conform to international conventions. Subjects received payment for participation in the experiment.

Exclusionary criteria to participate were: a history of neurological or otologic disease, significant abnormalities on neurological examination, a binocular visual acuity with corrective lenses worse than 20/40 in both eyes, abnormal, age-corrected audiometric function, abnormal ocular motor testing, i.e., spontaneous or gaze-evoked nystagmus, saccadic dysmetria, abnormal ocular pursuit, or abnormal optokinetic nystagmus, or abnormal vestibular function defined by a significant persistent positional nystagmus, abnormal caloric responses, or a significant directional preponderance (asymmetry) on earth-vertical axis rotation. Based on these criteria, 47 older and six young subjects were excluded. 19 of the 47 excluded older subjects had incomplete data largely a result of nausea during off-vertical axis rotation. Nine older and one younger subject were excluded because of abnormal hearing. Seven older subjects were excluded because of abnormal vestibular testing.

2.2. Information-processing tasks

There were two types of auditory information processing tasks used in this study: a non-spatial frequency choice reaction time task (FCRT) and a spatial lateralization choice reaction time task (LCRT). Subjects were seated in a rotational chair, and held a response button in each hand, with the instruction to place their hands on top of their thighs. The thumb of each hand was placed on top of the button, poised to press it. For the non-spatial task, tones were presented binaurally to the subject through headphones. These tones could be either low frequency (560 Hz) or high frequency (980 Hz), at an intensity of 80 dB SPL. These frequencies were chosen based upon tabled values of center frequencies of critical bands [19]. One of two tones was randomly assigned to be the “target tone.” Each subject was asked to push the button in the dominant hand in response to hearing the target tone, and to push the other button, in response to the non-target tone. The number of presentations for the target tone and the number of presentations for the non-target tones were equal. For the spatial task, a tone was presented to either the left or the right ear. Each subject was asked to push the button in the hand on the same side as the ear to which the stimulus was presented. The number of presentations to the right ear and the number of presentations to the left ear were equal. Subjects were instructed to react accurately and as quickly as possible to the tones; subjects were encouraged to respond quickly even if that resulted in making some mistakes. However, they were also advised to keep the number of mistakes low. Each subject performed four blocks of 32 trials for each of the two tasks during a non-test day that preceded the two test days. During training and testing a qualified technician monitored the data acquisition and assured that error levels remained between 5 and 10%.

2.3. Visual-vestibular conditions

Seven visual, vestibular, and visual–vestibular conditions were used. (Table 1) The vestibular-only conditions were accomplished by sinusoidal rotation of subjects in the dark about an earth-vertical axis at 0.05 Hz with a peak velocity of 50°/s (EVAR-SINE), rotation at a constant velocity about an off-vertical axis at 60°/s with a tilt angle of 30° (OVAR-CV), or rotation at a sinusoidal velocity about an off-vertical axis with a tilt angle of 30° at a frequency of 0.05 Hz with a peak velocity of 50°/s (OVAR-SINE). For the OVAR-CV condition, subjects were always initially rotated for 60 s about an earth-vertical axis to allow semicircular canal responses to subside [6]. Half of the subjects were rotated clockwise and the other half were rotated counterclockwise, with the rotation direction randomly assigned. For OVAR-SINE, subjects were tilted, then rotated. For the vision conditions, subjects were rotated about an earth-vertical axis at 0.05 Hz with a peak velocity of 50°/s while observing earth-fixed stationary floor-to-ceiling vertical stripes (VVR-SINE), sat in the stationary rotational chair while observing floor-to-ceiling vertical optokinetic stripes moving at a constant horizontal velocity of 60°/s (OKN-CV), or sat in the stationary rotational chair while observing floor-to-ceiling vertical optokinetic stripes moving horizontally at a sinusoidal velocity of 0.05 Hz with a peak velocity of 60°/s (OKN-SINE). A baseline condition consisted of no rotation (seated while stationary) in darkness. Eye position was monitored during testing with electrooculography to assure that subjects’ average eye position was straight ahead.

Table 1.

Sensory stimuli during visual-vestibular conditions

Condition	Semicircular canals	Otolith organs	Vision
EVAR-SINE	+	Not stimulated	Not stimulated
OVAR-CV	Not stimulated	+	Not stimulated
OVAR-SINE	+	+	Not stimulated
VVR-SINE	+	Not stimulated	+
OKN-CV	Not stimulated	Not stimulated	+
OKN-SINE	Not stimulated	Not stimulated	+
NO	Not stimulated	Not stimulated	Not stimulated

EVAR-SINE: sinusoidal earth vertical axis rotation in darkness; OVAR-CV: constant velocity off-vertical axis rotation in darkness; OVAR-SINE: sinusoidal off-vertical axis rotation in darkness; VVR-SINE: sinusoidal earth vertical axis rotation with a lighted visual surround; OKN-CV: constant velocity optokinetic stimulation; OKN-SINE: sinusoidal optokinetic stimulation in the dark; NO: stationary in darkness.

2.4. Protocol

On each of two test days, participants performed one of the two information-processing tasks. This experimental design was chosen to prevent confusion with respect to the cognitive task demands. The assignment of the particular information-processing task to each test day was randomized across subjects. The information processing task for each day was performed during a baseline (no rotation) condition and each of the visual vestibular conditions, leading to a total of seven task condition combinations for each day. The seven conditions (Table 1) were presented twice by grouping them into two blocks of the seven conditions; the order of presentation of the seven conditions was randomized within each block. Thus, over the course of the experiment, each subject underwent each of the 14 combinations of information-processing task and visual vestibular condition twice.

For each task-condition combination, 32 reaction time stimuli were presented to the subject. The auditory stimuli were presented soon after the trial commenced except for the OVAR-CV condition, for which auditory stimuli were presented only after the subject was tilted off-vertical. Stimuli were presented at a random time interval between 1.5–4.0 s, with an average time interval between trials of 2.75 s. Therefore, each task condition combination trial required about 90 s. Reaction times were recorded by computer with a resolution of 1 ms.

2.5. Data analysis

For the non-spatial task, the reaction time responses for the target and non-target tones were pooled for each trial. For the spatial task, the reaction time responses for the right ear and left ear presentations were pooled for each trial. Only reaction times from correct responses were processed. Reaction times that were less than 200 ms or greater than 1500 ms were removed. The median reaction time for each trial for each subject was computed and the average of the medians for the trials in each of the two blocks for each condition was then used for further analysis. An analysis of baseline median reaction times for effects of reaction time task type and age group used a repeated-measures ANOVA with task type as the within-subject variable, age group as the between-subject variable, and the interaction term. Subsequent analyses examining vestibular condition as well as reaction time task type employed a ‘task cost’ variable. Median reaction time for the baseline (no rotation in darkness) condition for each subject for each task type was subtracted from the median reaction time for each of the six stimulus conditions for that subject for that task type to yield task cost. A one sample t-test was used to assess whether task cost for each task-condition combination was non-zero. A repeated-measures ANOVA was performed to assess the effects of task, condition, age group, and their interactions on task cost. We controlled for gender by in cluding gender in the statistical model as a main effect without interaction terms. The full model was reduced by a stepwise elimination of non-significant interaction terms. Post-hoc comparisons were made for significant multiple-level variables. A significance level of 0.05 was used for the statistical analyses.

3. Results

Baseline reaction time responses taken in the dark without rotations for the non-spatial task (468 ms; s.e. 15 ms) and spatial task (389 ms; s.e. 15 ms) were significantly different (p < 0.001). There were no significant differences in the baseline median reaction times between the young (spatial task: 393 ms; s.e. 17 ms and non-spatial task 472 ms; s.e. 21 ms) and older group (spatial task: 386 ms; s.e. 17 ms and nonspatial task 465 ms; s.e. 20 ms), nor any interaction between age group and task.

The influence of age, task, and visual-vestibular condition on task cost was determined using repeated measures ANOVA. Task cost was influenced by the visual vestibular condition (p < 0.0001), and an interaction between condition and task: Condition*Task (p = 0.04) (Fig. 1). There was no significant main effect of age group, nor any interactions effects with age. Post hoc comparisons were made to explore how common characteristics across certain visual-vestibular conditions influenced task cost. Otolithic vs. semicircular canal influences were explored by comparing OVARSINE and OVAR-CV to the earth-vertical rotation condition EVR-SINE (Fig. 2a). Task cost for EVA-Sine was smaller than that for both OVAR-Sine (p <0.0001) and OVAR-CV (p < 0.0001) but OVAR-SINE and OVAR-CV were not significantly different from each other (p = 0.50). Visual vs. semicircular canal influences were explored by comparing VVR-SINE and OKN-SINE with EVA-SINE (Fig. 2b). Task cost for EVA-SINE was significantly smaller than that for both VVR-Sine (p < 0.001) and OKN-Sine (p < 0.001) but OKN-SINE and VVR-Sine were not significantly different from each other (p = 0.07). Thus, otolithic stimuli and visual stimuli had a larger task cost than semicircular canal stimulation alone. Combining semicircular canal with otolithic (OVAR-SINE) or combining semicircular canal with visual stimulation (VVR-SINE) had no significant additive effect.

Fig. 1.

Change in auditory reaction time from baseline (task cost) for six visual-vestibular conditions and two reaction time tasks. Data are pooled across age groups. Note that a significant change in reaction time from baseline occurs for all conditions and is largest for otolithic stimuli i.e. OVAR-SINE and OVAR-CV. Note that OVAR conditions show a lower task cost for the spatial task compared to the non-spatial task i.e. LCRT compared to FCRT, while the other four conditions show a greater task cost for the spatial task. Error bars represent one standard error. EVAR-SINE: Sinusoidal earth-vertical axis rotation in the dark; OKN-CV: Constant velocity optokinetic nystagmus; VVR-SINE: Sinusoidal earth-vertical axis rotation in the light; OKN-SINE: Sinusoidal optokinetic nystagmus; OVAR-SINE: Sinusoidal off-vertical axis rotation in the dark; OKN-CV: Constant velocity off-vertical axis rotation in the dark; FCRT-frequency choice reaction time task (non-spatial task); LCRT: lateralization choice reaction time task (spatial task).

Fig. 2.

Task costs for specific visual-vestibular conditions: a) Comparison of otolithic (OVAR-CV and OVAR-SINE) with non-otolithic (EVAR-SINE) vestibular conditions; b) Comparison of vision (OKN-SINE and VVR-SINE) with nonvision (EVA-Sine) conditions. Error bars represent one standard error. Abbreviations as in Fig. 1.

The interaction of condition and task was seen as a lower task cost for the spatial task during OVAR-SINE and OVAR-CV and a higher task cost for the spatial task in all the other conditions (Fig. 1). This difference suggested that there may be an overall visual influence on task cost. We evaluated this observation further by repeated measures ANOVA with independent variables being Task, Visual Condition, and their interaction. Visual Condition was categorized as those conditions without visual inputs (EVA-SINE, OVAR-SINE, OVAR-CV) and those with visual inputs (VVR-SINE, OKN-SINE, OKN-CV). Results showed no significant main effects of Visual Condition or Task, but a strong interaction effect (p = 0.006). Task cost was greater for the spatial task compared to the non-spatial task during the conditions with vision and not different for the conditions without vision (Fig. 3).

Fig. 3.

Task cost for vision conditions versus non-vision conditions by task type collapsing across age group. Note that the effect of task type on interference was different for the vision vs. no-vision conditions.

4. Discussion

Visual-vestibular stimulation has been shown to interfere with auditory information-processing tasks [5, 17]. The present study found a greater effect of otolithic stimulation than that found previously [5]. This result is thought to be based upon the higher intensity of otolithic stimulation used in this study compared to that used in our previous study. The present study extended our findings further by showing that visuo-ocular stimuli had a larger effect than semicircular canal stimulation. Once again, we found that age group did not significantly affect task cost for interference between a cognitive task and visual-vestibular stimulation despite including significantly older individuals than those tested previously. This study also assessed the importance of task type and found that task type interacted with condition such that there was larger interference for the spatial vs. the non-spatial auditory processing task for conditions that included a visual stimulus.

Our previous research has suggested that interference between visual-vestibular stimulation and auditory information processing occurs as sensory stimuli compete for processing rather than at later, relatively motoric stages of processing [5,15]. The neurophysiologic basis for the interference between visual-vestibular stimulation and auditory information processing may be related to cortical-cortical inhibition since functional imaging studies have shown that vestibular stimulation can inhibit cortical activation induced by visual stimulation [1,3,9] and vestibular stimulation can reduce cortical excitability in visual processing areas [14].

Our results suggest that otolithic stimulation has a larger effect on information processing than either visual or semicircular canal stimulation. This finding may be based on the unusual character of the otolithic stimulation that we used, i.e., off-vertical axis rotation, which, for constant velocity rotations, represents a pure otolithic and thus somewhat unusual vestibular stimulus. Possibly, this type of stimulation has a large task cost because vestibular signals are relayed to the cortex prominently via the parabrachial nucleus, which is especially excited by novel vestibular stimuli [10]. The larger influence of visual conditions vs. semicircular canal stimulation may be related to powerful cortical cortical inhibition expected from a moving visual stimulus based on functional imaging studies using optokinetic stimuli [9]. Our finding that task cost for combined visual and semicircular canal stimuli did not exceed that for visual stimulation alone and that the task cost for combined otolithic and semicircular canal stimuli did not exceed that for otolithic stimulation alone suggests a non-linear, i.e., non-additive, task cost for these stimuli. This nonlinear behavior may reflect the complex dynamics of cortical-cortical inhibition [2]. A limitation of this study is that the “strength” of the different visual-vestibular conditions may not have been uniform based on properties such as stimulus velocity rather than on which sensors, i.e. semicircular canal vs. visual vs. otolithic, were stimulated.

The absence of an effect of age on task cost may have been the result of insufficiently demanding tasks. This concept is supported by the absence of an effect of age on baseline measures of RT. An alternative explanation for the absence of an effect of age is that decreased connectivity [18] with advanced age may have lead to less interference if interference is based in part on sensory sensory inhibition at the cortical level. However, a limited capacity for information processing in the older subjects [11] may have led to increased interference that offset a decrease in connectivity.

This study showed a relatively weak but significant interaction between task type and condition such that task cost was larger for the spatial task for stimuli that included vision but not for conditions that excluded vision. This result suggests that the visual stimuli used in this study engaged spatial orientation circuitry preferentially. A caveat regarding this result is that the visual and non-visual conditions being compared in Fig. 3 included stimuli other than simply vision vs. no vision. For example, the visual conditions included optokinetic stimulation whereas the non-visual conditions included otolithic stimulation. The higher task cost for the spatial task during vision conditions cannot be explained based on task difficulty as the non-spatial task is actually more difficult based on single task reaction times. Moreover, the increased spatial task cost found in the vision conditions is unlikely to be related simply to the presence of a visual stimulus as compared with darkness. Our prior research has indicated that for subjects who were maintaining upright stature on stable and moving surfaces, the presence of vision reduces rather than increases task cost [4]. Also, the stimulus with the highest task cost overall i.e., off-vertical axis rotation, was associated with a higher task cost for the non-spatial stimulus. Possibly, the task cost for the complex and unusual otolithic stimulation produced by off-vertical axis rotation was based primarily on non-specific interference with attentional mechanisms rather than on interference with spatial orientation pathways per se. Notably, our results differed somewhat from those of Gresty et al. [7] who reported that earth-vertical axis rotation interferes with spatial but not verbal tasks. Our results also differ from those of Gresty et al. [8], who reports that Coriolis stimulation, which is quite provocative for motion sickness similar to OVAR, degraded spatial task performance preferentially. The basis for these differences in results probably relates to differences in how the tasks were performed. Although both our studies and those of Gresty et al. [7,8] used tasks that contained a spatial component, Gresty et al. [8] report a type of auditory Stroop-like test and Gresty et al. [7] report a spatial transformation task.

Our results may have implications for patients with vestibular disorders, whose vestibular responses are abnormal and thus could potentially experience more interference than that found in normal individuals. A prior study suggested increased interference between cognitive processing and visual-vestibular stimulation in patients with well-compensated unilateral peripheral vestibular abnormalities [5]. Additional studies will be required to confirm and extend this result in patients, particularly regarding how spatial vs. non-spatial tasks would interfere with balance. Possibly, patients undergoing vestibular rehabilitation may benefit from an assessment and treatment component that includes dual task paradigms to uncover and remediate excessive interference between vestibular stimulation and information processing that could increase the risk of falling.