Combined EEG and Brain-Perfusion-SPECT Measurements in Cochlear-Implant Users: Networks of Speech Understanding - OMAI 2018- The 10th International Symposium on Objective Measures in Auditory Implants

Irina Schierholz ^1,2 Mariella Kessler ^2,3 Martin Mamach ^2,4 Florian Wilke ⁴ Anja Hahne ⁵ Lilli Geworski ⁴ Frank M. Bengel ³ Georg Berding ^2,3 Pascale Sandmann ⁶

¹Department of Otorhinolaryngology, Hannover Medical School, Hannover, Germany
²Cluster of Excellence, "Hearing4all", Germany
³Department of Nuclear Medicine, Hannover Medical School, Hannover, Germany
⁴Department of Medical Physics and Radiation Protection, Hannover Medical School, Hannover, Germany
⁵Saxonian Cochlear Implant Center, Department of Otorhinolaryngology, Technical University Dresden, Dresden, Germany
⁶Department of Otorhinolaryngology, University of Cologne, Cologne, Germany

Background: Speech perception with cochlear implants (CIs) is highly variable.

Objectives: We combined brain-perfusion-SPECT and EEG to elucidate the underlying mechanisms.

Methods: 14 CI-users (58±11y) performed a speech-discrimination task with semantically correct or incorrect sentences while a 96-channel-EEG was recorded. After 2min. 740MBq ^99mTc-HMPAO were injected and a dual-head-SPECT scan was acquired 1.5h p.i. (“speech condition”). A second scan was performed at rest (“rest condition”). SPM8 was used to compare both conditions and correlate their difference to EEG (N100/N400), speech audiometry (HSM-sentence-test), and cognitive (MWT-B/size-comparison-span-test) data. Subgroups of good and poor CI-users were created by a median-split procedure (HSM in noise: poor<48%; good>48%).

Results: Speech processing resulted in bilateral recruitment of auditory regions (Brodmann areas (BAs) 21/22/41/42; p<0.001). An enhanced N400, reflecting the detection of semantic violation, was related to stronger activation of left-sided temporal areas (BA20; p<0.01). Better working-memory and linguistic competence were related to stronger activations (“speech condition”) in auditory (BA41/42/22/21), Broca’s (BA44/45), left frontal (BA8) and premotor areas (BA6; all p<0.001). Bad performers showed stronger activation of Broca’s (BA45; p<0.01), left parietal (BA40/2; p<0.001), frontal (BA8; p<0.01), prefrontal (BA9/46; p<0.01) and left premotor (BA6; p<0.01) areas, whereas good performers showed stronger activation of auditory (BA21/left; p<0.001; BA21/right;BA22/left; p<0.01) and temporal regions (BA20/38; p<0.01).

Conclusion: Combined SPECT/EEG measurements in the context of a speech-discrimination task revealed recruitment of a temporo-frontal network and a relation between SPECT/EEG, speech and cognitive measures. Good and poor CI-users showed different cortical activation patterns. Results might help to understand/improve auditory rehabilitation in CI-users.