niels_round
Niels Henrik Pontoppidan

Research Area Manager/ Research Engineer

> Learn more

mail@eriksholm.com

Historically the audiogram describes the individual hearing loss. Professor Brian Moore's group at Cambridge University developed the TFS1 test to improve hearing aid fittings in line with individual TFS capabilities. Their findings - and Eriksholm's replication hereof - indicate that regardless of hearing loss, listeners are insensitive to temporal fine structure.

Testing for TFS sensitivity

The TFS1 test was developed by Professor Brian Moore's group at Cambridge University to assess individual sensitivity to temporal fine structure (TFS)1. We decided to replicate an earlier version to gain experience with TFS testing and verify their early conclusions2.

The listeners heard two intervals of four complex tones in noise. One interval consisted of four reference tones with the same pitch, while the other interval interleaved two reference tones with two tones with slightly higher pitch. Thus the second interval should be heard as changing.

Stimuli

The tones mainly consisted of five components with f0 Hz between neighbouring components. The reference tones centred at N⋅f0 Hz are harmonic. The shifted tones centred at N⋅f0+Δf Hz are non harmonic. Due to the constant frequency spacing, the beats between components - and thus the temporal envelope - remained constant, regardless of whether the tone was a reference or shifted tone.

If only the five aforementioned components were present, the shifted tones could evoke excitation pattern cues compared to the reference tones. Consequently the TFS-1 test stimuli also had flanking tonal components (e.g. at 800, 700 Hz, … and 1400, 1500 Hz,… in Figure 2) with increasingly lower amplitude, to maintain the excitation pattern difference between the reference and shifted beeps below the normal-hearing listeners' discrimination threshold

13044

Results

19 listeners with mild-to-moderate hearing impairment, and 8 normal-hearing listeners participated. While the normal-hearing listeners found the task rather easy, their hearing-impaired counterparts were unable to discriminate the reference and shifted tones. Since the temporal envelope was constant and the excitation pattern difference negligible, this replicated the previous findings1,2 that hearing impairment also reduces sensitivity to TFS, thereby preventing discrimination of simple sounds.

Figure 2: Sketched spectrogram of stimuli having f0=100 Hz, N=11, and Δf=50 Hz. This is the main condition where components are believed to be unresolved for both normal and hearing-impaired listeners. The grey background indicates that the testing occurs in noise.

Temporal or spectral fine structure

We concluded that discrimination was due to TFS abilities. This has been challenged by arguments that the excitation pattern difference was not negligible, and that the normal-hearing listeners performance was due to sensing the spectral shift4.


Further reading

1 Moore BCJ, Sek A (2009) Development of a fast method for determining sensitivity to temporal fine structure, International Journal of Audiology, 48(4), p. 161-171.

2 Hopkins K, Moore BCJ (2007). Moderate cochlear hearing loss leads to a reduced ability to use temporal fine structure information. Journal of the Acoustical Society of America, 122(2), p. 1055-1068.

3 Oxenham AJ, Micheyl C, Keebler M (2009). Can temporal fine structure represent the fundamental frequency of unresolved harmonics? Journal of the Acoustical Society of America, 125(4), p. 2189-2199.

4 Santurette S (2011). Neural coding and perception of pitch in the normal and impaired human auditory system, Ph.D. dissertation, Department of Electrical Engineering, Technical University of Denmark.

Learn More

  • learnmore-poster

    Poster

    The TFS1 test reveals mild hearing loss