Changes in the response properties of inferior colliculus neurons relating to tinnitus

Abstract

Tinnitus is often identified in animal models by using the gap prepulse inhibition of acoustic startle. Impaired gap detection following acoustic over-exposure (AOE) is thought to be caused by tinnitus "filling in" the gap, thus, reducing its salience. This presumably involves altered perception, and could conceivably be caused by changes at the level of the neocortex, i.e., cortical reorganization. Alternatively, reduced gap detection ability might reflect poorer temporal processing in the brainstem, caused by AOE; in which case, impaired gap detection would not be a reliable indicator of tinnitus. We tested the latter hypothesis by examining gap detection in inferior colliculus (IC) neurons following AOE. Seven of nine unilaterally noise-exposed guinea pigs exhibited behavioral evidence of tinnitus. In these tinnitus animals, neural gap detection thresholds (GDTs) in the IC significantly increased in response to broadband noise stimuli, but not to pure tones or narrow-band noise. In addition, when IC neurons were sub-divided according to temporal response profile (onset vs. sustained firing patterns), we found a significant increase in the proportion of onset-type responses after AOE. Importantly, however, GDTs were still considerably shorter than gap durations commonly used in objective behavioral tests for tinnitus. These data indicate that the neural changes observed in the IC are insufficient to explain deficits in behavioral gap detection that are commonly attributed to tinnitus. The subtle changes in IC neuron response profiles following AOE warrant further investigation.

Keywords: acoustic over-exposure; auditory; behavior; electrophysiology; gap detection; response types; tinnitus; tinnitus animal model.

Figure 1

Behavioral signs of tinnitus. (A) Behavioral performance after AOE is shown for a representative “tinnitus” animal, expressed as change in PPI, i.e., a ratio of performance before vs. after noise exposure, at each background frequency. Values <1 indicate poorer gap detection, while a value close to 1 indicates no effect of noise exposure. Black bars indicate frequencies with significantly worse gap detection (P < 0.05). (B) The number of animals exhibiting gap detection deficits at each background frequency. Note: some animals exhibited gap detection deficits at more than one frequency. (C) There was no significant correlation between behavioral performance and ABR threshold shifts. Solid line indicated linear regression. Dotted lines show 95% confidence intervals. Each point represents the behavioral performance for a given background condition in a single animal, relative to the ABR threshold shift at the corresponding frequency.

Figure 2

Neuronal activity in the inferior colliculus. Representative examples are shown of onset (A), sustained (B), and offset (C) single-units recorded in the IC. PSTH plots were generated in response to pure tone gap detection stimuli incorporating gaps of varying duration 4, 8, 10, 20, 50, and 75 ms are shown. Red bars indicate the duration of the first 200 ms stimulus, followed by a variable gap, and finally the second 50 ms stimulus.

Figure 3

Changes in the proportion of IC response profiles. (A) The proportion of single-units exhibiting either an onset or sustained-response profile is shown for each experimental group. Offset units have been excluded from this graph for the purposes of clarity (due to their low incidence). Onset responses significantly increased (relative to sustained units) in no tinnitus and tinnitus animals, compared with controls (***P < 0.0001). (B) Number of units recorded for each experimental group, separated according to response-type classification. The number of units recorded in no tinnitus GPs was considerably lower than controls and tinnitus GPs. (C) The increase in the proportion of onset units in tinnitus GPs was isolated to the contralateral (relative to AOE) IC (***P < 0.0001). In no tinnitus GPs, this increase was present in both the IC_ipsi and the IC_contra.

Figure 4

Neural gap detection in the IC. (A) Mean MGDTs in response to pure tone stimuli at CF for control (n = 96), no tinnitus (n = 42), and tinnitus (n = 109) animals. (B) Mean MGDTs in response to BBN for control (n = 76), no tinnitus (n = 42), and tinnitus (n = 103) groups (*P < 0.05). (C) Mean MGDTs in response to NBN for control (n = 43), no tinnitus (n = 25), and tinnitus (n = 33) GPs. (D) The percentage of gap-detecting cells, relative to total cell count in response to pure tone stimuli for control, no tinnitus, and tinnitus GPs. For each gap duration (1, 2, 4, 8, 10, 20, 50, and 75 ms) the percentage of cells able to detect a gap less than or equal to a given duration is shown. (E) The percentage of gap-detecting cells with a BBN carrier. (F) The percentage of gap-detecting cells with an NBN carrier. (G) Distribution of the number of units with MGDTs for each gap duration in response to pure tone stimuli, shown for control and tinnitus GPs. For the purposes of clarity, no tinnitus GPs have been omitted. (H) Distribution of MGDTs in response to BBN stimuli. (I) Distribution of MGDTs in response to NBN stimuli. (J) The averaged population responses of IC neurons in control (black line), no tinnitus (dark gray line), and tinnitus (light gray line) groups for gaps of 4 ms duration, in response to a pure tone carrier. (K) Averaged population responses to gaps of 8 ms duration. (L) Averaged population responses to gaps of 50 ms duration.

Figure 5

Hemispheric differences in gap detection for pure tones. (A) Mean MGDTs in response to pure tones are shown when data were sub-divided according to recording side. The percentage of gap-detecting cells are also shown for left IC (B) and right IC (C) in response to pure tones.

Figure 6

Hemispheric differences in gap detection for BBN. (A) Mean MGDTs with a BBN carrier from control, no tinnitus, and tinnitus animals, sub-divided according to recording side (**P < 0.01). The percentage of gap-detecting cells for left IC (B) and right IC (C) are also shown.

Figure 7

Neural GDTs in the different types of IC neurons. Mean MGDTs are shown for onset and sustained units (offset units were excluded owing to a small sample size) in control, no tinnitus, and tinnitus GPs.