We have developed a new finite element (FE) model of human right ear, including the accurate geometry of middle ear ossicles, external ear canal, tympanic cavity, and mastoid cavity. The FE model would be suitable to study the dynamic behaviors of pathological middle ear conditions, including changes of stapedial ligament stiffness, tensor tympani ligament (TTL), and tympanic membrane (TM) stiffness and thickness. Increasing stiffness of stapedial ligament has substantial effect on stapes footplate movement, especially at low frequencies, but less effect on umbo movement. Softer TTL will result in increasing umbo and stapes footplate displacement, especially at low frequencies (f < 1000 Hz). When the TTL was detached, the vibration amplitude of umbo increased by 6 dB at 600 Hz and two peaks (300 and 600 Hz) were found in the vibration amplitude of stapes footplate. Increasing the stiffness of tensor tympani resulted in a slightly decreased umbo amplitude at very low frequencies (f < 500 Hz) and significantly decreased displacement up to 12 dB at middle frequencies (1000 Hz < f < 4000 Hz). However, the amplitude change of stapes footplate is very sensitive to the TTL stiffness, especially at low frequency (f < 1000 Hz). The increased stiffness of TM resulted in reduced umbo and stapes footplate displacement at frequencies <1500 Hz and increased displacement at frequencies >1500 Hz. As (TM) thickness was increased, the umbo displacement was reduced, especially at very low frequencies (f < 600 Hz). Otherwise, the stapes displacement was reduced at all frequencies.

"Correlation of speech and music appreciation in post-lingually deaf Combi 40/40+ cochlear implant users." Cochlear Implants International, 4(sup1), pp. 68–69

Speech tests have been performed on 6 subjects for comparing the standard 12-channel continuous interleaved sampling (CIS) strategy (CIS12), the 7-channel CIS strategy (CIS7) and the 7-of-12 strategy in the MED-EL COMBI 40+ system. An ABAB experimental design was used whereby each strategy was reversed and replicated. Speech tests were performed in quiet (vowels, consonants, monosyllables, sentences) and noise (sentences). Results showed that for vowels, CIS12 is significantly superior to CIS7, for consonants and sentences CIS12, CIS7 and 7-of-12 performed equally well, and that for monosyllables 7-of-12 is significantly superior to both CIS12 and CIS7. In addition, 7-of-12 is superior to CIS7 by almost the same amount as CIS12, but in this case the difference is not significant. Further, all strategies have been found to be equally robust in noise with respect to sentence understanding. The differences between CIS12 and 7-of-12 on the one hand and CIS7 on the other hand may be attributed to decreased spectral resolution of the latter. The fact that – in contrast to what has been reported for the SPEAK strategy – 7-of-12 is equally robust in noise as the CIS strategies is explained by the use of higher stimulation rates, wider frequency bands and a higher percentage of channels stimulated in each cycle.