Abnormal resting-state cortical coupling in chronic tinnitus

Abstract

Background: Subjective tinnitus is characterized by an auditory phantom perception in the absence of any physical sound source. Consequently, in a quiet environment, tinnitus patients differ from control participants because they constantly perceive a sound whereas controls do not. We hypothesized that this difference is expressed by differential activation of distributed cortical networks.

Results: The analysis was based on a sample of 41 participants: 21 patients with chronic tinnitus and 20 healthy control participants. To investigate the architecture of these networks, we used phase locking analysis in the 1-90 Hz frequency range of a minute of resting-state MEG recording. We found: 1) For tinnitus patients: A significant decrease of inter-areal coupling in the alpha (9-12 Hz) band and an increase of inter-areal coupling in the 48-54 Hz gamma frequency range relative to the control group. 2) For both groups: an inverse relationship (r = -.71) of the alpha and gamma network coupling. 3) A discrimination of 83% between the patient and the control group based on the alpha and gamma networks. 4) An effect of manifestation on the distribution of the gamma network: In patients with a tinnitus history of less than 4 years, the left temporal cortex was predominant in the gamma network whereas in patients with tinnitus duration of more than 4 years, the gamma network was more widely distributed including more frontal and parietal regions.

Conclusion: In the here presented data set we found strong support for an alteration of long-range coupling in tinnitus. Long-range coupling in the alpha frequency band was decreased for tinnitus patients while long-range gamma coupling was increased. These changes discriminate well between tinnitus and control participants. We propose a tinnitus model that integrates this finding in the current knowledge about tinnitus. Furthermore we discuss the impact of this finding to tinnitus therapies using Transcranial Magnetic Stimulation (TMS).

Figure 1

Phase locking averaged over all connections. Differences between the tinnitus and the control group were significant in the alpha 9–12 Hz and the gamma 48–54 Hz frequency range. Asterisks mark significant frequency bins after Bonferroni correction.

Figure 2

Schematic display of the alpha and the gamma network. Connections with a significant group difference were plotted as edges in the networks. The nodes were named by: LF = Left Frontal, RF = Right Frontal, LT = Left Temporal, RT = Right Temporal, LP = Left Parietal, RP = Right Parietal, ACC = Anterior Cingulate Cortex, and PCC = Posterior Cingulate Cortex.

Figure 3

Negative relationship between the alpha and the gamma network. As an indicator for the strength of the alpha and the gamma network we calculated the sum of all phase locking values for both frequency bands. Decoupling of long-range alpha is associated with an increase of long-range gamma coupling. Individuals with tinnitus can be separated from controls with an accuracy of 83%.

Figure 4

Schematic display of alpha and gamma networks for both tinnitus subgroups with tinnitus of short and long duration. LF = Left Frontal, RF = Right Frontal, LT = Left Temporal, RT = Right Temporal, LP = Left Parietal, RP = Right Parietal, ACC = Anterior Cingulate Cortex, and PCC = Posterior Cingulate Cortex.

Figure 5

Centrality of the left temporal cortex in the gamma network. As a measure for the centrality of the left temporal cortex we calculated the „weighted degree" for both subgroups by adding the number of links that connect to the left temporal source, weighted by the strength of the connections. The centrality of the left temporal cortex was compared with the centrality of other regions. Error bars indicate the standard error of the mean over the respective tinnitus group.

Figure 6

Relationship between the tinnitus duration and the long-range gamma coupling of the left temporal cortex (r = -.56, p < .01, figure 6a) and the remaining sources (r = .74, p < .001, figure 6b).