An Electrophysiological Version of the Audible Contrast Threshold Test

The audible contrast threshold (ACT) test is a behavioral spectro-temporal modulation detection test used to approximate speech-in-noise performance. This study focused on the clinical reliability and validity of an electrophysiological version of the ACT test (E-ACT). The primary objectives were to evaluate the test-retest reliability of the E-ACT and explore its relationship with the behavioral ACT test and aided speech-in-noise performance. Since the E-ACT employs stimuli based on a tonal carrier while the ACT uses a noise carrier, a secondary objective was to investigate the behavioral correlation between performance with noise-carrier and tonal-carrier stimuli. We hypothesized that there would be a significant relationship between the E-ACT and the behavioral ACT.

The E-ACT utilizes the auditory change complex, a cortical response to suprathreshold sound changes, elicited here by changes between spectro-temporal modulated targets and non-modulated references. The 4-Hz temporal and two-cycles/octave spectral modulation rates were identical to those used in the ACT test. Two experiments were conducted to evaluate the E-ACT: The first experiment (experiment A) was conducted in a university setting with 18 adult participants. The second experiment (experiment B) was conducted in a hospital setting with 47 hearing-impaired adult participants. Both experiments included ACT and E-ACT measurements. In addition, experiment A incorporated behavioral tonal-ACT measurements, while experiment B included assessments of speech-in-noise performance. Electroencephalogram data were analyzed using the objective Fmpi (individualized multi-point F test) detector to determine the presence of responses.

The E-ACT test and retest were significantly correlated ( R2 = 0.62, p < 0.001). The E-ACT values correlated significantly with ACT values ( R2 = 0.18, p < 0.001), with most participants showing better thresholds for ACT compared with E-ACT. However, unexpectedly, some participants had worse ACT values compared with E-ACT. E-ACT values also correlated with aided speech-in-noise performance ( R2 = 0.09, p = 0.04). Last, ACT and tonal-ACT values were significantly correlated ( R2 = 0.53, p < 0.001), with thresholds obtained using tonal-ACT averaging 1.6 dB lower than those obtained using ACT.

The E-ACT provides clinicians with an objective discrimination measure that can be utilized with patients unable to participate in behavioral tests. Both E-ACT and ACT offer insights into suprathreshold hearing, which can supplement the audiogram.