The Vestibular Implant Input Interacts with Residual Natural Function

Abstract

Objective: Patients with bilateral vestibulopathy (BV) can still have residual "natural" function. This might interact with "artificial" vestibular implant input (VI-input). When fluctuating, it could lead to vertigo attacks. Main objective was to investigate how "artificial" VI-input is integrated with residual "natural" input by the central vestibular system. This, to explore (1) whether misalignment in the response of "artificial" VI-input is sufficiently counteracted by well-aligned residual "natural" input and (2) whether "artificial" VI-input is able to influence and counteract the response to residual "natural" input, to show feasibility of a "vestibular pacemaker."

Materials and methods: Five vestibular electrodes in four BV patients implanted with a VI were available. This involved electrodes with a predominantly horizontal response and electrodes with a predominantly vertical response. Responses to predominantly horizontal residual "natural" input and predominantly horizontal and vertical "artificial" VI-input were separately measured first. Then, inputs were combined in conditions where both would hypothetically collaborate or counteract. In each condition, subjects were subjected to 60 cycles of sinusoidal stimulation presented at 1 Hz. Gain, asymmetry, phase and angle of eye responses were calculated. Signal averaging was performed.

Results: Combining residual "natural" input and "artificial" VI-input resulted in an interaction in which characteristics of the resulting eye movement responses could significantly differ from those observed when responses were measured for each input separately (p < 0.0013). In the total eye response, inputs with a stronger vector magnitude seemed to have stronger weights than inputs with a lower vector magnitude, in a non-linear combination. Misalignment in the response of "artificial" VI-input was not sufficiently counteracted by well-aligned residual "natural" input. "Artificial" VI-input was able to significantly influence and counteract the response to residual "natural" input.

Conclusion: In the acute phase of VI-activation, residual "natural" input and "artificial" VI-input interact to generate eye movement responses in a non-linear fashion. This implies that different stimulation paradigms and more complex signal processing strategies will be required unless the brain is able to optimally combine both sources of information after adaptation during chronic use. Next to this, these findings could pave the way for using the VI as "vestibular pacemaker."

Keywords: bilateral vestibular areflexia; bilateral vestibulopathy; neural prosthesis; vestibular implant; vestibular prosthesis; vestibulo-ocular reflex.

Figures

Figure 1

Polar plot illustrating the angles of eye movements. A horizontal eye movement to the right corresponded to an angle of 0o (blue arrow), a horizontal eye movement to the left to an angle of 180o (orange arrow). A completely vertical eye movement upwards corresponded to an angle of 90o (purple arrow) and a completely vertical eye movement downwards to an angle of 270o (green arrow).

Figure 2

Plots presenting for all subjects the raw eye movement signals of each condition in the horizontal and vertical planes. Positive horizontal velocities correspond to movements to the right and negative horizontal velocities to movements to the left. Positive vertical velocities correspond to movements upwards and negative vertical velocities to movements downwards.

Figure 3

Averaged eye and head movement signals of each condition in the horizontal and vertical planes. Positive horizontal velocities correspond to movements to the right and negative horizontal velocities to movements to the left. Positive vertical velocities correspond to movements upwards and negative vertical velocities to movements downwards. Note that since the electrically evoked vestibulo-ocular-reflex (eVOR) condition involved no head movements, a hypothetical horizontal head movement is plotted corresponding to the electrical stimulus of the vestibular implant. The amount of cycles measured is given, as well as the number of cycles available for analysis after data cleaning.

Figure 4

Vectors of peak total eye velocities in the excitatory and inhibitory phases of stimulation, plotted for each electrode in each condition [vestibulo-ocular-reflex (VOR), electrically evoked vestibulo-ocular-reflex (eVOR), totalVOR+, and totalVOR−]. The gain is represented by the vector magnitude. The angle of the response is represented by the vector angle (according to the polar plot in Figure 1). Dots represent the medians, gray bars the 95% confidence intervals and the open bars the interquartile ranges of the vectors of peak total eye velocities. Red represents peak total eye velocities obtained during the excitatory phases of stimulation, blue during the inhibitory phases. The amount of analyzed excitatory and inhibitory peak total eye velocities is given, as well as the amount of peaks available for analysis after data cleaning. Note that, to improve visibility, the scale of the polar plots for each subject was optimized for individual responses and consequently is not uniform across subjects.

Figure 5

Asymmetry of the eye movement responses plotted for each electrode in each condition [vestibulo-ocular-reflex (VOR), electrically evoked vestibulo-ocular-reflex (eVOR), totalVOR+, and totalVOR−]. The widest part of each diamond represents the median, upper and lower parts of the diamond the 95% confidence intervals, and the bars the interquartile ranges. The amount of analyzed excitatory and inhibitory peak total eye velocities is the same for each electrode and condition as in Figure 4.

Figure 6

Phases of the eye movement responses plotted for each electrode in each condition [vestibulo-ocular-reflex (VOR), electrically evoked vestibulo-ocular-reflex (eVOR), totalVOR+, and totalVOR−]. Zero corresponds with “in phase,” ±180° with “counter phase.” Positive values correspond with a phase lead, negative values with a phase lag. The middle bar represents the median, the gray area the 95% confidence interval and the outer bars the interquartile ranges. Depending on the case, horizontal, vertical, or both phases are presented. The amount of cycles measured is given, as well as the amount of cycles available for analysis after data cleaning.

Figure 7

Schematic representation of the interaction between residual “natural” input and “artificial” vestibular implant input in the excitatory and inhibitory phases of the totalVOR+ and totalVOR− conditions of all electrodes. Arrows show the median vector of peak total eye velocities obtained during excitatory and inhibitory phases of the experimental trials. Excitatory vectors contain a red color; inhibitory vectors contain a blue color. Vestibulo-ocular-reflex (VOR) is represented by a plain arrow, electrically evoked vestibulo-ocular-reflex (eVOR) by a striped arrow, and totalVOR by a green edged arrow. In the excitatory phases of totalVOR+, the excitatory phases of VOR and eVOR were combined. In the excitatory phases of totalVOR−, the excitatory phases of VOR were combined with the inhibitory phases of eVOR, since the gyroscopes were inversed during the totalVOR− condition. In the inhibitory phases of totalVOR+, the inhibitory phases of VOR and eVOR were combined. In the inhibitory phases of totalVOR−, the inhibitory phases of VOR were combined with the excitatory phases of eVOR.