Corrigendum to “Diagnosing Noise-Induced Hearing Loss Sustained During Military Service Using Deep Neural Networks”

Moore, B. C. J. , & Schlittenlacher, J. (2023). Diagnosing Noise-Induced Hearing Loss Sustained During Military Service Using Deep Neural Networks. Trends in Hearing, 27. doi:10.1177/23312165231184982

The author of this paper contacted us informing that the Table 2 of this paper contained three copy and paste errors. The correct value for the weight from input HTL right 1 kHz to hidden unit 1 is −0.036. The correct value for the weight from input AAHL 2 kHz to hidden unit 1 is −0.333. The correct value for the weight from input AAHL 8 kHz to hidden unit 2 is −0.102.

These errors had only a very small influence on the outcome. With the wrong values, the sensitivity was 0.979 rather than 0.986 with the correct values. The specificity was unaffected by the errors, being 0.902.

The article has been updated to reflect this correction.

Table 2.

Weights of MLP(18), which had 18 input features.

	From input	To hidden unit 1	To hidden unit 2	From	To output
HTL right	1 kHz	−0.036	0.129	Unit 1	1.598
	2 kHz	0.113	0.213	Unit 2	1.011
	3 kHz	0.061	0.236	Bias	0.656
	4 kHz	0.470	0.376
	6 kHz	0.421	0.136
	8 kHz	−0.337	0.097
HTL left	1 kHz	0.689	0.368
	2 kHz	−0.009	−0.090
	3 kHz	−0.127	0.079
	4 kHz	0.436	0.621
	6 kHz	0.507	0.136
	8 kHz	−0.009	0.035
AAHL	1 kHz	0.224	−0.138
	2 kHz	−0.333	0.079
	3 kHz	−0.239	−0.189
	4 kHz	−0.065	−0.014
	6 kHz	0.064	0.097
	8 kHz	0.126	−0.102
	Bias	0.811	0.941