Training Cadaver Dogs on Pig Remains May Lead to Evidentiary Challenges

Perhaps the most important research being conducted on cadaver dogs is that of Mary Cablk of the Desert Research Institute in Reno and her colleagues at the University of Nevada-Reno.  In June 2011, we discussed an important study in which she tested the ability of dogs to find teeth in a natural environment. 

In a paper appearing in Forensic Science International, Dr. Cablk, Erin Szelagowski, and John Sagabiel analyzed the volatile organic compounds (VOCs) in the headspace above partially decomposed animal tissue samples and compared the chemicals found with what could be expected from human decomposition.  The animal remains were from cows, pigs, and chickens.  The researchers found that the “VOC signatures” of the animal remains differed from each other, but also from humans. 

The difference between pigs and humans was particularly important to describe because pig remains are sometimes used to train human remains detection dogs—cadaver dogs.  See, for instance, the training instructions of Southwest Rescue Dogs, Inc., which state:

“SRDI dogs train on real cadaver material. Alternative sources are available, such as pseudoscent. Pig flesh is similar to human and is used as a training aid if you do not have access to cadaver material. SRDI also uses soil from a burn area for training aids.”

This practice must now be questioned because the researchers concluded that pig VOC signatures shared only seven of 30 human-specific compounds, while nine unique compounds in pig VOC signatures were not found in human samples.  Curiously, chicken and human samples were actually closer to the human signature. 

HRD Dogs

The researchers note that compared to drug and bomb dogs, scientific literature on human remains detection (HRD) dog capabilities is sparse.  They note (citing Lorenzo et al. 2003), as they had earlier in the paper about finding human teeth, that the belief that HRD dogs can differentiate between human and animal remains “has yet to be tested in a rigorous experimental framework.” 

They also note that certifying HRD dog teams often requires that dogs not indicate on animal remains placed on the ground surface, citing procedures of the International Police Work Dog Association, the North American Police Work Dog Association, and the California Emergency Management Agency.  (It should perhaps be noted that FEMA has put its imprimatur on IPWDA and other organizations for human remains canine search teams.)   As to such testing procedures, the researchers state:

“[S]uch demonstrations cannot be considered proof that the HRD dog team does not alert on animal remains because these demonstrations lack a blind or double-blind protocol stipulation, thus enabling bias from handler, evaluator, or others.”

For this statement the researchers discuss the article by U.C. Davis researchers on handler beliefs, discussed by us in a draft paper posted on the website of the Michigan State University College of Law.

The researchers also state that the “potential for the handler to see the animal remains during the test is likewise problematic.”  No specific citation for this sentence is provided, but some organizations create testing environments where this possibility does exist.  The International Police Work Dog Association provides that animal remains “may be on the surface to act as a distraction.”  The NASAR Canine Certification Program (revised December 31, 2010), provides that distractions that may be used during testing include “commercially manufactured animal urine, horse or cow manure, legally obtained skeletal or other remains of animals, animal skins obtained and possessed within local wildlife regulations, or animal antlers obtained and possessed within local wildlife regulations.”

Training with Human Remains

The researchers note that training “rarely involves the use of entire decedents because such aids are difficult to access, store, and maintain as well as having significant legal issues to possess.”  In some countries use of human remains for such a purpose would be illegal.  Dogs may be trained on fabric or soil that has been in direct contact with human remains or with small tissue samples.  The researchers also note that “immediately after death a corpse begins the process of decomposition and different parts of the body decompose at different rates allowing for certain compounds to emanate from the body faster than others.”  They state (citing Rebmann 2000) that often “the level of decomposition of the training aid and the purity of the sample is unknown.”

The researchers observe that dating “the relationship between VOCs from human remains and the odor signature that canines recognize as target odor is poorly understood.”  An instrument may detect VOCs that a dog does not.  Consequently, studies determining ratios of compounds using solid-phase microextraction-gas chromatograph/mass spectrometry (SPME-GC/MS), according to the researchers, “may not be realistic to the actual odor signature.” 

Cross-Contamination

The researchers state that cross-contamination of training aids occurs in two ways:

“First, cross-contamination occurs when a training aid of one type (e.g., bone) is stored in proximity to a training aid of another type (e.g., blood) in permeable containers and when training aids are stored together in the same container without physical separation. Either situation results in training aids that no longer emit an odor signature true to the original target, rather the odor emitted becomes a mixture from multiple targets…. Training aids also become contaminated with permeable container odor, such as plastics like polyvinyl chloride, which degrades the desired human remains specificity training on the dog.”

Cross-contamination can also arise from inadequate storage, improper handling and placement, poor choices of packaging materials, and vapor pressure, storage temperature, or humidity. Another problem is that training aids are stored in aerobic conditions but a cadaver may be buried in anaerobic conditions (citing Vass et al. 2008).

Pseudo Scents

Pseudo scents for heroin, marijuana, and explosives have been designed to match measured VOC profiles, yet one study (Marcias et al. 2008) found that dogs did not respond to the pseudo scents.  “The lack of response by the trained detector dogs indicated a mis-match between the instrument-measured, human-interpreted profile and the dogs’ recognized scent signature.” (The Supreme Court is poised to decide whether a drug dog’s training and certification is insufficient to provide probable cause for a vehicle search in Florida v. Harris (Docket No. 11-817), a case on which the Court granted certiorari on March 26.  If a dog has been trained using a fallible technique, could it be said to be “well-trained?”  

Although a certification procedure is something of a threshold to assure the dog has some skills, an analysis of the dog’s performance history would more fully establish a canine team’s capabilities.)

The researchers state that if “synthetic training aids were available then the reliability of HRD dogs might increase simply through improved specificity and sensitivity to the trained target odor.”  It is not clear if the researchers were aware of, or chose not to mention, that there already are pseudo scents, such as Sigma-Aldrich’s Pseudo Corpse Scent Formulations.  The adoption of pseudo scents would require many standards to be modified to permit use of such scents. 

What is Uniquely Deceased Human in an Odor Signature?

The researchers used as a model a prior study (Macias et al. 2009) that sought to suggest possible target signature compounds that could be tested for recognition by HRD dogs.  Cablk and her colleagues wanted to compare VOCs in the headspace of cows, chickens, and pig remains with those of the 2009 study on human remains “to examine potential compounds that might be part of the odor signature that is uniquely ‘deceased human’ and identify possible generic decomposition indicators.”  The current study replicated the 2009 study in part to eliminate potentially contributing VOC sources such as soil microbial activity and nearby air emission sources. (As already noted, Southwest Rescue Dogs, Inc. permits use of soil from a burn area as a training aid.)

If dogs can differentiate between the scent of human decomposition and that of animal decomposition, the VOCs found in animal decomposition should differ from those found in human decomposition enough for the dog to recognize the difference.  Alternatively, the ratios of specific VOCs could allow the dogs to differentiate.  The researchers found that there are, in fact, different VOCs involved in human decomposition and the three species of animals that were used.  Curiously, however, particularly considering the preference for using pigs as substitutes for humans in some forensic procedures, the researchers found that by all measures, the compounds shared by humans and pigs were least similar. 

Vass et al. (2008) listed 30 compounds as key markers of human decompositions, of which seven were found in animal samples by Cablk and her colleagues.  These compounds could not be considered human-specific.  The researchers also discussed some compounds that might be found in some people but not others because of environmental factors.  It is also to be noted that while dogs have much more sensitive olfactory systems than we do, there are some chemicals as to which their sensitivity may not be significantly different from ours (Tonosaki and Tucker, 1985). 

Future Research

The researchers state that a further study should involve presenting HRD dogs “with human samples and animal samples in a scientific framework to determine if the canines alert on both human and animal scent or demonstrate the ability to differentiate between the two by alerting on human scent only. This information might allow us to eliminate compounds or at least guide in determining what compounds or ratios of compounds might be need to be included in a canine’s training aid.”  Also, instrument detectors being developed may ultimately replace dogs in cadaver searches. 

Future research should perhaps take into account that untrained dogs sometimes find human remains and training methods such as two-step procedures whereby (1) the dog is first trained to indicate to an indirect stimulus (pig) and (2) then undergoes extinction training to disassociate decomposing pig odor from cadaver odor.  Headspace studies may not readily incorporate an understanding of such approaches. 

Conclusion

Other groups are looking at the VOC profile produced by pig carcasses and human remains, so additional studies may be expected. 

As we noted in Police and Military Dogs, the work of cadaver dogs is relatively infrequently questioned in court because a cadaver is found and identified by DNA, dental comparison, or other procedures.  The work of the dog becomes historical to the case, and does not receive more than a brief mention in a judicial opinion.  Nevertheless, this study may have an impact on future prosecutions where a body has not been found and the dog’s alert is taken as proof that the suspected decedent was present at a location, or in cases where the body has been found and the prosecution’s problem is to establish that it was in the defendant’s car or house or at some other specific location (consider the recent Casey Anthony prosecution involving a car trunk).  If the defense can make a plausible argument that a dead pig or other animal might have been at the location, and establish that the dog’s training aids included swine remains, a court may have to consider the implications of this research. 

Cablk, M.E., Szelagowski, E.E., and Sagebiel, J.C. (2012). Characterization of the Volatile Organic Compounds Present in the Headspace of Decomposing Animals Remains, and Compared with Human Remains.  Forensic Science International (2012) (doi:10.1016/j.forsciint.2012.02.007).

      This blog was written by John Ensminger and L.E. Papet. 

      

      Additional Sources: 

1. Hoffman, E., Curran, A., Dulgerian, N., Stockham, R., and Eckenrode, B. (2009) Characterization of the Volatile Organic Compounds Present in the Headspace of Decomposing human remains. Forensic Science International, 186, 6–13. 
2.  Lit, L., Schweitzer, J., and Oberbauer, A. (2011). Handler Beliefs Affect Scent Detection Dog Outcomes. Animal Cognition, 14, 387–394. 
3.  Lorenzo, N., Wan, T., Harper, R., Hsu, Y., Chow, M., and Furton, K. (2003). Laboratory and Field Experiments Used to Identify Canis lupus var. familiaris Active Odor Signature Chemicals from Drugs, Explosives, and Humans. Analytical and Bioanalytical Chemistry. 376, 1212–1224. 
4.  Macias, M., Harper, R., and Furton, K. (2009). A Comparison of Real Versus Simulated Contraband VOCs for Reliable Detector Dog Training Using SPME-GC–MS. American Biotechnology Laboratory, 26, 26–27. 
5. Rebmann, A., David, E., and Sorg, M. (2000). Cadaver Dog Handbook. CRC Press, Boca Raton.
6. Tonosaki, K., and Tucker, D. (1985). Responsiveness of the Olfactory Receptor Cells in Dog to Some Odors. Comparative Biochemistry and Physiology 81A, 7–13. 
7.  Vass, A., Smith, R., Thompson, C., Burnett, M., Dulgerian, N., and Eckenrode, B. (2008). Odor Analysis of Decomposing Buried Human Remains.  Journal of Forensic Sciences, 53, 384–391.