Table 1 Results of linear regression models asking what types of syllables were most easily discriminated.

From: Discrimination of natural acoustic variation in vocal signals

What features of syllables explain performance differences in syllable rendition discrimination?

 

χ2

Adj P-value

R2

Fixed effect

FundFreq

3.77

0.224

0.152

MeanFreq

9.27

0.023*

0.294

PeakFreq

4.13

0.224

0.170

FreqMod

1.40

0.630

0.091

AmpMod

2.95

0.322

0.192

Entropy

1.65

0.598

0.101

Contt

0.03

0.993

0.032

Contf

0.67

0.993

0.175

Interaction

Duration:Motif

12.81

0.023*

0.262

Harmonicity:Motif

14.56

0.021*

0.265

  1. Each model consisted of mean PC* scores on each task other than the whole motif renditions (4 syllable rendition tasks for each of the 3 motifs) for each subject as the response variable (37 observations), the fixed effects were an acoustic feature of the relevant background syllable (e.g. duration), motif set (3 levels), and the interaction between motif set and acoustic feature, and the random effects were subject and task (12 levels) to account for the repeated measures design of the experiment. The formulas for the models were as follows: PC* ~ Feature*Motif + (1|Subject) + (1|Task). Chi-square (χ2) values (comparing the full model against a model that includes all other terms) are given for the fixed effects involving an acoustic feature, unless the interaction between acoustic feature and motif set was significant, in which case that is reported. P-values for chi-square tests were adjusted using the Benjamini–Hochberg false discovery rate procedure. Marginal R2 is given for the fixed effects for each model. * < 0.05.
  2. FundFreq fundamental frequency, MeanFreq mean frequency, PeakFreq peak frequency, FreqMod frequency modulation, AmpMod amplitude modulation, Contt spectral continuity over time, Contf spectral continuity over frequency.
Back to article page