Do I smell a rat (er, mouse) in TSA’s future?

I hope so.

Israel21C recently reported on new technology that might someday replace invasive pat-downs and body scans TSA treats us to at the airport:

Israeli startup Bioexplorers has developed a new and unique way to sniff out terrorists – literally. After years of research, company CEO Eran Lumbroso tells ISRAEL21c, Bioexplorers has hit upon a foolproof, non-invasive and easy method to detect contraband in purses, luggage and even cargo – using mice.

Like dogs, mice have an excellent sense of smell and they’re relatively easy to train. As much as it pains to me admit it, mice offer some advantages. Their small size means they’re cheaper and easier to keep than dogs are, and because they don’t need human handlers, mice also won’t be sensitive to their prejudices.

The proposed system will combine low tech mice with high-tech training and screening equipment. The target to be screened will move through a passageway equipped with fans that extract the air surrounding the target and deliver it to chambers containing several mice. Each mouse is trained to respond to a single odor. When a mouse detects the target odor it moves into a second chamber and sets off an alarm.

Bioexplorers’ system is interesting because not only could it allow for less invasive screening than existing measures but, as an improvement over existing canine detection methods, it could also give screeners the ability to determine exactly which odorants have been detected.

Initial tests are promising.

[Bioexplorers] has conducted several tests at sites in Israel to ensure that the sensors work in real situations, including at Tel Aviv’s Azrieli Mall. More than 1,000 people passed through a Bioexplorers sensor – some having been given “suspicious” objects and substances to hold – and the mice made the right call every time, says Lumbroso.

The company says they can train mice in just two weeks using a patented Skinnerian computer program. Mice will be expected to work four hour shifts with eight hour rest periods and each mouse’s career is anticipated to last about two years.

It’s an interesting idea and there may be additional uses for the tiny detectors. Bioexplorers’ representatives say they’re also working on systems designed for medical use.

They may be able to detect drugs and explosives at minute concentrations, but the mouse’s nose may be too refined for use in the wine industry. The io9 forums report that researchers at Japan’s Hiroshima University conducted an experiment in 2008 to see if they could teach mice to tell different kinds of wine apart. The experiment was successful so the group decided to take the idea a step further and see if they could teach mice to tell different brands of red wine apart.

The results were interesting.

Only two of the ten mice tested displayed the ability to consistently tell the red wines apart. Six others performed on a what appeared to be a purely random basis.

But the final pair of connoisseur mice could not be persuaded to respond the target wine. These mice were consistently drawn to a specific non-reward wine indicating that they preferred the smell of this wine to the food rewards offered by the alternate choice.

I’m not terribly surprised to hear that well-fed laboratory mice will sometimes prefer an alluring scent over certain food rewards. Smell plays a huge part in the social and intellectual life of a mouse. A good smell, or an interesting one, may provide the kind of intellectual stimulation that could be more lacking in a laboratory mouse’s life than food is.

And while mice may replace dogs in situations where a static location like an airport or freight terminal needs to be screened for a wide array of compounds, I suspect that dogs will continue to be the detector of choice in field situations for some time to come.

See no evil. Read no evil. Cite no evil.

The internet hosts hundreds of articles warning you about the dangers of electronic training collars (e-collars). Ruth over at Spot Check recently summarized a few of the most often cited studies in a post on the heated rhetoric surrounding the recent ban on the use of e-collars in Wales. Her post was the inspiration for this one.

The literature is full of references to studies by Schalke et al., Schilder and van der Borg and more recently, Herron et al. whose authors warn us that e-collar training (and indeed, any use of aversives) is unpleasant, painful, frightening — and pointlessly ineffective.

If you spend some time reviewing these articles, as I recently did, you might assume that no research supporting the use of e-collars is currently available.

And you’d be wrong.

Given the widespread references /cites to studies that support the idea that e-collars are not only cruel and abusive, but that they can also elicit aggressive behavior — imagine my surprise when I came across an article providing strong evidence that e-collars were astonishingly effective in rehabilitating aggression in dogs.

Daniel F. Tortora’s study, titled “Safety Training: The Elimination of Avoidance-Motivated Aggression in Dogs,” was published in the Journal of Experimental Psychology: General in 1983. The article is only available by purchase but is well worth $11.95 if you have an interest in this area. (Note: the article was also published in Australian Veterinary Practitioner in 1984, 14 (2), 70–74.)

Tortora took an elegantly simple approach to treating what he referred to as “avoidance-motivated aggression”. He proposed that because avoidance-motivated aggression is learned and maintained as an avoidance response, the most effective way to counter-condition it would be to teach the dogs nonaggressive avoidance responses.

Tortora defines avoidance-motivated aggression as “a form of instrumental aggression that involves attacks or threats of attack directed toward one or more of the dog’s human caretakers”. Avoidance aggression typically starts out as aggressive avoidance responses to things like physical discomfort (such as from grooming), intrusions on areas that the dog views as his territory and commands he doesn’t want to comply with. According to Tortora, these dogs usually suffer from a lack of training and predictability in their lives and therefore feel like they lack control over their environment.  They behave like they expect bad things to happen and the only way to prevent the bad things is through aggression. When their frustrated owners resort to after-the-fact punishment, the dog’s expectations are reinforced, a feedback loop is created and the dog’s aggression escalates.

Tortora’s proposed remedy for this common, dangerous and difficult to remedy form of aggression consisted of teaching the dogs “nonaggressive, prosocial habits” such as AKC’s CDX level obedience exercises. He predicted that the probability of post-training aggressive behavior would be inversely proportional to the number of obedience exercises a dog gained proficiency in. The program also included teaching the dogs a conditioned safety signal that was used to reinforce good behavior and build the dogs’ confidence.

All exercises were introduced with the slip collar, then e-collar training was overlayed onto the introductory work. The e-collars used could emit two different tones, and tones and stimulation could be delivered separately or in conjunction with each other. The dogs were trained to perform 15 different commands at increasing levels of difficulty. These included: stand, down, come, go, hold, drop, sit, off, place, fetch, in, stay, play, no, heel, and hup. As commands were mastered, they were practiced in environments of increasing distraction. The dogs were initially trained by experienced trainers (Tortora doesn’t describe their qualifications but all were apparently able to train the dogs to a minimum of CDX level around significant distractions) in a board and train environment. Once the dogs were able to consistently perform the exercises under distraction without the e-collar, training was transferred to their owners, who used the e-collar only as needed to proof exercises.

Tortora stated that the dogs could be safely returned to their owners because: “Safety training with companion dogs, however, produces changes of long duration, perhaps even permanent changes. These changes in behavior readily transfer readily from the trainer to the dog’s owners and others.”

Many people are concerned that the stress of e-collar training will make dogs fearful or aggressive. While the dogs developed an initial conditioned anticipatory fear reaction during the escape training portion of Tortora’s program, their fear was extinguished during the subsequent avoidance and proofing stages. Upon reviewing these results, Tortora stated “It seems that the impact of safety reinforcement is to make the dog less fearful generally and better able to withstand trauma.”

How effective was this work? Well, in the abstract Tortora states that the program:

… resulted in complete and permanent elimination of aggression in all of the 36 dogs tested. In addition, it produced extremely extinction-resistant prosocial avoidance responses, significant increases in the dogs’ emotional stability, an avoidance-learning and safety acquisition response set, and improvements in measures of the dogs’ “carriage.”

Take a few minutes to let that sink in. If a study demonstrated similar results for clicker or food lure training it would be cited on tens of thousands of sites across the internet. The author would be the darling of popular dog magazines and a regular presenter at dog training conferences. Heck, I bet he’d even have his own television show – and (unlike another popular television dog trainer) there wouldn’t be a torch and pitchfork mob out to lynch him.

While I understand that the literature can be (and often is) cherry-picked to support preconceived notions even in peer-reviewed studies, I am absolutely stunned by the dog world’s shunning of Tortora’s work. His article is very rarely cited in recent studies related to ecollars, aversives, dog training and aggression — and when it is, it is not unusual for him to be misquoted or taken out of context. (details on that below the break)

Given the outstanding success Tortora had in rehabilitating aggressive dogs and the fact that his article appeared in a well-known journal published by the American Psychological Association, why are studies published by Schalke, Schindler and Herron (and opinion pieces written by Pat Miller) touted as landmark studies on e-collar use while his work languishes in anonymity?

Using e-collars to train dogs is a controversial and emotionally-charged issue. This is largely because, as Steven Lindsay writes:

… the word shock is loaded with biased connotations, images of convulsive spasms and burns, and implications associated with extreme physical pain, emotional trauma, physiological collapse, and laboratory abuses.

Shock scares us. Despite the fact that electrical stimulation can now be used to relieve pain, most people simply cannot come to terms with the idea that a ‘shock’ can be used as anything but a terrifying and harshly punitive bolt from god.

Unlike those commonly in use today, early electronic training collars could only  be used in a harshly punitive way – and much of the laboratory research that has been done on shock, aversion, escape and avoidance was horrifyingly cruel. Along with the strongly negative connotations associated with the word “shock”, the ugly history of the use of shock in behavior modification studies also affects our feelings and opinions about its place in dog training.

The current literature on the use of aversives (especially electronic ones) in dog training shows a striking lack of articles that present results that call popular ideas favoring positive reinforcement only dog training into question. And unfortunately, as we recently saw in Wales, the results published in these studies are being used to further a political agenda.

There are far too many cases where great scientific advances were made based on a piece of odd, apparently anomalous or unpopular bit of work that could very easily have fallen by the wayside. Rejecting, ignoring or suppressing data and ideas that don’t fit in with popular thought is a dangerous kind of censorship. And it is crucial that we do all we can to it in a world where science has an increasingly important effect on the personal and regulatory decisions we make.

————————————————————————–

Below the break: Links and brief summaries of recent literature related to using e-collars to train dogs, and some notes on the journals the articles are published in.

Well — This Explains it

If you are foolish enough to read too much of what is currently published about animal psychology, animal rights, operant conditioning, dominance hierarchies, raw food diets, titers versus vaccines, the use of corrections in dog training, breed specific legislation, theories of mind, aromatherapy, early spay/neuter, evolution and whether or not your dog really will resent you for putting that silly costume on him at Halloween — you’re probably at least a little bit confused by the rabidly opinionated and utterly contradictory information you’ve found.

I’m one of those morons who reads too much.  And having spent far too much of my life absorbed with books, laboratory data and computer modeling – I decided to allay a bit of my own confusion by (what else) doing a bit of research on the net. 

Eureka!  I’ve found the explanation.  The flow chart below was recently featured on The Lounge of the Lab Lemming.  And it explains everything.

And, if you insist on getting serious about being able to read an article, study or paper critically — check out A Magical Journey Through the Land of Logical Fallacies  (here’s part 2) and this short, but insightul article on Ethics and Peer Review at nature.com

Fur-Kids or Kindred Spirits?

Humans have always felt a strong kinship with animals.  Our ancestors believed that animals were aware and that their lives and communications were as meaningful as ours.  They were better able than we are to find meaning in subtle nuances of animal behavior because their survival depended on it – and because they saw animals as kindred spirits who understood the world in a different way than they did. 

Today most humans live in an environment far removed from the one our ancestors’ relied on for their survival, and most of us find it more difficult to recognize and understand the deeper meaning underlying animal behavior than our ancestors did.  We still feel a strong connection to animals, but changes in our environment and cultures have drastically changed the way we experience that connection.  Instead of seeing animals as kindred spirits with different perspectives and ways of life, many of us now think of them as amusing copies of our human selves.  Animals as kindred spirits, guides, teachers and partners have given way to fur-kids wearing designer dog coats.

This common, excessively anthropomorphized view may be part of the reason why many behaviorists and sociobiologists believe that assigning human, or human-like emotions and intentions to animals is a scientific taboo (as it has been since the time of Descartes).  But in a recent reversal, instead of denying the similarities between us, some scientists are studying how the emotions, modes of communication and motivations of animals do, in many ways, resemble our own.

The problem is not that we anthropomorphize, but that we tend to do it in the wrong ways. 

If we view anthropomorphism as a means, rather than an end, and use it to study the ways that animal behavior resembles human behavior we may gain valuable insights.  Research on the qualities that we share with them may help us better understand the areas where their perspectives differ from ours.  If, on the other hand, we adopt anthropomorphism as an end in itself, we simply stop at assigning human values and motivations to animals.  And when we do this, not only do we lose an incredible opportunity to expand our horizons, but we condemn them to a life and a set of expectations that they can never meet or be fully content with.

“So many baffling aspect of animal behavior are like that – baffling only because we fail to appreciate that the animal’s range of senses is not the same as our own: different but not always inferior.”  Hans Brick “The Nature of the Beast”