The rigor mortis development of large Atlantic salmon was evaluated at
a commercial plant using a sensory method, a mechanical rigorometer a
nd low-frequency vibrations. The fish were classified according to rig
or state using a neural network. Assessments were done with fish subje
cted to two different levels of handling stress prior to slaughter. Wh
ite muscle high-energy phosphates and inosine monophosphate were used
as indices of stress showing that handling stress may be considerable
during commercial slaughter. Results indicate rigor assessment should
be nonde structure and classification of rigor state was possible usin
g a low-frequency vibration method combined with a neural network.