Research Article

Modeling of Auditory Neuron Response Thresholds with Cochlear Implants

Table 1

Parameter estimates of the linear mixed model evaluating the relation between neural response thresholds and demographical, clinical, and technical factors.

Fixed effectUnivariate analysisMultivariate analysis
Crude -estimate (SE) valueAdjusted -estimate (SE) value

Age at test0.13 (0.13)0.310.29 (0.19)0.11
Age of profound deafness0.06 (0.12)0.58
Duration of profound deafness−0.54 (0.48)0.26−0.17 (0.64)0.78
Age at implantation0.03 (0.11)0.77
Impedance * impedance −0.03 (0.01)0.09−0.11 (0.02)<0.01
Distance * distance3.43 (1.44)0.022.20 (1.46)0.13
Neuron frequencyFigure 6(a)<0.01Figure 6(b)<0.01
Position
Scala vestibulirefref
Scala tympani−0.21 (1.99)0.91−8.50 (1.97)<0.01
Electrode array
Slim Straight refref
Perimodiolar −5.11 (12.95)0.69−5.94 (16.26)0.71
Straight1.02 (17.72)17.7213.43 (22.07)0.54
Etiology of hearing loss
Progressive SNHLrefref
Viral0.53 (17.61)0.972.55 (21.24)0.90
Genetic−5.36 (6.24)0.39−3.44 (7.22)0.63
Autoimmune−8.22 (9.74)0.40−10.50 (11.94)0.37
Ototoxic−19.34 (9.79)0.04−24.33 (13.24)0.06
Otosclerosis−2.23 (11.00)0.83−4.09 (13.45)0.76

SE: standard error.
-estimate: mean increase in NRT threshold according to the increase of one unit of the considered quantitative explanatory variable. For qualitative explanatory variables, the -estimate is the mean difference in NRT compared to the reference category (ref).
-estimates were compared to the value 0; a corresponding value <0.05 indicates a significant association between NRT and the explanatory variable.
“Age of profound deafness” and “Age at implantation” were not entered into the multivariate model because of their high colinearity with “Age at test.”