Analysis of DLF measurements.

A repeated-measures ANOVA was first performed, with frequency and run number as within-subject factors. No significant difference across frequencies or across runs was found before rehabilitation (for frequency, F(8,128)=1.501, p=0.175; for run, F(2,128)=0.375, p=0.693), at 1 month (for frequency, F(8,128)=0.417, p=0.907; for run, F(2,128)=1.135, p=0.346), at 3 months (for frequency, F(8,128)=0.430, p=0.899; for run, F(2,128)=2.115, p=0.153) or at 6 months (for frequency, F(8,128)=0.604, p=0.712; for run, F(2,128)=0.872, p=0.437) post-rehabilitation.

For each subject, the best DLF was located in the vicinity of the Fc, between Fc-3/8 octave and Fc+1/4 octave. The analysis centred on the individual best DLF (bDLF).

Figure 1 shows the individual alteration in DLF values during auditory rehabilitation in the nine subjects. At bDLF, all subjects except one (subject M.M.) had better DLFs before auditory rehabilitation than one month after, one subject had better DLFs at 3 months than before (subject J.P.B.), and 2 subjects had better DLFs at six months than before (subjects J.P.B. and P.C). The frequencies corresponding to Fc and bDLF for each subject can be found in Table 1.

The bDLF is located in the vicinity of the hearing-loss cut-off frequency. DLFs are plotted on a log scale as a percentage of the frequency tested.

To compare DLF thresholds, a two-way repeated-measures ANOVA was then performed for the various test times (before, 1 month, 3 months and 6 months after rehabilitation) and for various frequencies around bDLF (Fref, bDLF–1/8 octave, bDLF and bDLF+1/8 octave) (Figure 2). The interaction between ‘frequency’ and ‘time course of the rehabilitation’ proved significant (F(9,72)=3.214; p=0.003). Post-hoc tests revealed that DLF thresholds at bDLF were significantly lower than at bDLF+1/8 octave before hearing-aid fitting (p=0.008). DLF thresholds at bDLF were then found to increase during hearing-aid fitting: at 1 month and 3 months post-rehabilitation, DLF thresholds were significantly higher than before the rehabilitation (at 1 month, p=0.004; at 3 months, p=0.016). No statistical difference was found at 6 months post-rehabilitation for this frequency, and no alteration in DLF was found for the other frequencies during the auditory rehabilitation (Figure 2). No significant effect of ‘frequency‘ alone (F(3,72)=3.214, p=0.640) or ‘time course of the rehabilitation’ alone (F(3,72)=0.982, p=0.418) was found.

DLFs are measured at the best frequency limen (dFc)+/-1/8 octave and at Fref. DLFs are plotted on a log scale as a percentage of the frequency tested; the standard error of the mean is indicated for each mean DLF.