S: Sensory & Nutritive Qualities of Food JFS S: Sensory and Nutritive Qualities of Food Portable Electronic Nose for Detection of Spoiling Alaska Pink Salmon ( Oncorhynchus gorbuscha) J. CHANTARACHOTI, A.C.M. OLIVEIRA, B.H. HIMELBLOOM, C.A. CRAPO, AND D.G. MCLACHLAN ABSTRACT: The ability of a portable hand-held electronic nose (EN) in detecting spoilage of whole Alaska pink salmon (Oncorhynchus gorbuscha) stored at 14 ◦ C and in slush ice (1 ◦ C) was investigated. Fish were sampled daily at 14 ◦ C for up to 3 d, while fish stored in slush ice were sampled at various intervals up to 16 d. Sensory evaluations indicated that fish were rejected at day 3 when stored at 14 ◦ C and at day 12 when stored in slush ice. Aerobic bacteria counts for fish skin at 14 ◦ C ranged from 3.4 log 10 colony-forming units (CFU)/cm 2 (day 0) to 4.8 log 10 CFU/cm 2 (day 3) and for fish stored in slush ice ranged from 3.4 log 10 CFU/cm 2 (day 0) to 5.5 log 10 CFU/cm 2 (day 16). The correct classification rate using forward stepwise general discriminate analysis was 85% and 92% for EN analysis of belly cavity volatiles for fish held at 14 ◦ C and in slush ice, respectively. A predictive model may be developed for spoilage of whole Alaska pink salmon by analyzing belly cavity odors using the EN. Keywords: electronic nose, fish spoilage, pink salmon, seafood quality Introduction A major obstacle for the Alaska salmon industry has been the in- consistent quality of its raw material (Alaska Salmon Industry Task Force 2003). For this industry to remain competitive, it must improve quality and especially consistency. This requires proces- sors to search for novel technology that can provide the product consistency demanded by the consumers. Electronic nose (EN) is a relatively new technology that may serve as an inexpensive and objective quality control tool for the Alaska salmon processors. Re- search on the use of ENs for the evaluation of seafood quality showed that EN sensor responses can be correlated with sensory analysis re- sults for odor evaluation of a variety of fishery products such as cod, farmed Atlantic salmon, tilapia, catfish, yellow-fin tuna, redfish, and cold-smoked salmon (di Natale and others 1996; Luzuriaga and Bal- aban 1999; Du and others 2001; Korel and others 2001a, 2001b; Du and others 2002; Olafsdottir and others 2002; Olafsdottir and others 2005). The Cyranose 320 TM (Cyrano Sciences, Inc., Pasadena, Calif., U.S.A.) is a hand-held EN composed of 32 individual thin-film carbon-black polymer sensors. Recent reports on the use of this EN for determination of food quality show that the ability of this instru- ment to distinguish between sample categories is case dependent. Van Deventer and Mallikarjunan (2002) reported excellent discrim- ination, with a cross-validation value of 100%, for food packaging taints using the Cyranose 320. Three levels of retained solvents in a 2-layer polymer film, used for food packaging, were compared (Van Deventer and Mallikarjunan 2002). We have used the Cyranose 320 together with static headspace gas chromatography coupled to mass spectrometry analysis to try distinguishing degrees of skin wa- MS 20060104 Submitted 2/15/2006, Accepted 4/5/2006. Authors Chantara- choti, Oliveira, Himelbloom, and Crapo are with Fishery Industrial Tech- nology Center, Univ. of Alaska Fairbanks, 118 Trident Way, Kodiak, AK 99615. Author McLachlan is with Canadian Food Inspection Agency, P.O. Box 1060, Dartmouth, Nova Scotia, Canada B3B 1Y9. Direct inquiries to author Oliveira (E-mail: ffamo@uaf.edu ). termarking in canned pink salmon (Oliveira and others 2005). Subtle changes in the volatile patterns resulting from the different degrees of skin watermarking in salmon were difficult to distinguish by ei- ther method of analysis (Oliveira and others 2005). Apparently, the complex nature of food matrices seems to pose greater difficulties for the Cyranose 320 to distinguish between sample categories than for single chemical taints. The objective of this study was to test the ability of the Cyra- nose 320 in detecting spoilage of whole Alaska pink salmon (Oncorhynchus gorbuscha) stored at 14 ◦ C and in slush ice. The EN “sniffed” belly cavity odors and gill odors. Sensory evaluation and microbial analysis were conducted in conjunction with EN analysis. Materials and Methods Fish storage preparation Wild fresh whole pink salmon (O. gorbuscha) were collected from a seafood processing plant in Kodiak Island (Alaska) during summer 2004 and immediately brought to the Fishery Industrial Technology Center (FITC) pilot plant. Salmon used in this study were less than 24 h postmortem and were sampled the same day and from the same fishing vessel. All fish were caught in waters surrounding Ko- diak Island. Whole fish weight ranged from a minimum of 1.5 kg to a maximum of 2.5 kg. All salmon were graded on the Alaska Seafood Marketing Inst. skin color evaluation guides (ASMI 2004). One hun- dred and fifty fish used in this study were grade A and had no skin watermarking. Fish were separated into 2 groups; the first group was kept submerged in seawater in a temperature-controlled chamber (model 9030, VWR Scientific, West Chester, Pa., U.S.A.) operated at 14 ◦ C. The second group was kept in slurry of ice water averaging 1 ◦ C. Temperatures were monitored using an iButton temperature data logger (model DS1921G, Semiconductor/Maxim Corp., Dallas, Tex., U.S.A.). Fish stored at 14 ◦ C were analyzed daily for 3 d while those stored on slush ice were analyzed at various intervals up to 16 d of storage. S414 JOURNAL OF FOOD SCIENCE—Vol. 71, Nr. 5, 2006 C  2006 Institute of Food Technologists doi: 10.1111/j.1750-3841.2006.00050.x Further reproduction without permission is prohibited