Thursday, January 29, 2009

The Death of Due Process?

Check out this pitt-bull-ban story — as well as the earlier posts on the story. Whatever happened to reasonable search and seizure?

I’ll Be Damned

I can’t claim to be one of the major actors: But I was there, and part of the Greek chorus, when DKL LifeGuard began to come crashing down.

To set the scene, it was the quarterly meeting of the Appalachian Search & Rescue Conference’s board of directors at the University of Virginia, where our sister team the Blue Ridge Mountain Rescue Group is based. At this particular meeting, two guest speakers presented a rescue device they were selling: the DKL LifeGuard, an artifice for detecting the heartbeats of hidden and/or trapped disaster victims, lost persons, what have you. The device detected human, and only human, heartbeats, they said.

Today I can’t remember with certainty what tweaked my skeptical bone that day. I’ve heard too much of the ensuing story, learned too many amazing things about what this device was — or rather, wasn’t — to sort out the source of my own concern. But I do remember thinking that the way of using the device (a gun-like detector pivot-mounted on a pistol grip) [1], moving it in a straight line from side to side (without, mind you, making an arc), would inevitably keep its business end pointed at whatever you were facing.

It may be a little hard to believe, but that was how it worked: You pointed it at your suspected subject’s location, moved it side to side, and if a human heart beat behind a wall, or trees, or whatever, the thing would supposedly point at the source.

Something other than Newtonian mechanics lay behind my disquiet, though. While the head rep talked to us, his assistant was standing behind him, waving the damned thing at us like a long-haired 1970s magician. Without looking back, senior guy says, “Joe [or whatever his name was], quit it.”

A bell went off in my head; a whiff of snake oil (or maybe cat shit).

Well, if these guys had planned to dazzle the natives, they’d picked the wrong village. ASRC has always been a bit of a geek kingdom within search and rescue, and on that day we had some people there who were far more qualified than me to see the holes in the LifeGuard story. When, following the presentation, a group of us shared lunch, Keith Conover, MD, emergency physician and a founder of my own Allegheny Mountain Rescue Group as well as the ASRC itself, pointed out that, since animal and human EKGs are electromagnetically indistinguishable, the claim that the device can distinguish between them was ridiculous on the face of it. A lifelong HAM, he also pointed out that the length of the device’s antenna was wrong [2].

It got worse. Gene Harrison, another founder of the ASRC, was also there. I’m not sure Gene has a website; at the time, at least, he worked for Mitre, a nonprofit organization that advises Free World governments on technical issues, often at the highest levels of secrecy. Gene worked on communications in a manner that would have obligated him to kill us if he’d told us more. He, too, was buying none of it.

And they both, I believe, used the same word, a word many of us were thinking but didn’t dare voice:

FRAUD

Some of the folks at that little lunch were scandalized: not by the presentation, but by the baldness of that conclusion. I was among a third group, who shared grave doubts about the “technology” but shied from the harshness of the word, as apt as it turned out to be.

I won’t retell the whole story, because Keith does it so well and completely on his website; but subsequent revelations included an evaluation by Sandia National Laboratories that showed the damned things weren’t any better than random chance, and that the devices contained an open circuit bridged by a human hair — yes, an old dowsing standard. Keith was one of the heroes of the story, pushing this with the authorities (some of DKL’s defenders among whom turned out to have been early purchasers, possibly desperate to avoid embarrassment) and pushing back, quite effectively, when DKL’s lawyers threatened.

I’ve wanted to use the LifeGuard story for a while on this blog, but had resisted precisely because it’s such a high-powered context: You present any scientific finding next to it, and the concept of fraud hangs over the whole enterprise. I wanted to be very careful before loading anybody’s research with that kind of context.

Hadn’t occurred to me that I might use it in a completely opposite way: To frame the story of a scientific finding (two findings, in this case) that forced me to re-evaluate something I’d thought was hooey, but may in fact be absolutely legitimate.

I’m talking about aromatherapy, a hippie-dippy, loosey-goosey, on-the-face-of-it ludicrous alternative therapy that I’d thought was on par with homeopathy — but which, I am amazed to learn, may have some significant scientific backing [3].

In a slightly elderly (2005) report just indexed on MEDLINE (“MEDLINE Daily Update,” no less), Mamoru Tanida and crew from Osaka and Niigata universities report that rats who smelled essential oil of grapefruit — or its major component limonene — experienced an increase in blood pressure, a reduction in vagal gastric nerve activity, and an increase in nerve activity to the kidneys. The important thing here is that the Osakans could prevent all of the above effects with classical interventions for blocking the sense of smell, such as bathing the nasal cavity with zinc sulfate or surgically damaging the olfactory bulbs. So we can be pretty sure the effect is working through the sense of smell rather than, say, the limonene being absorbed and acting in the bloodstream or elsewhere.

Takakazu Oka and company from the University of Occupational and Environmental Health in Kitakyushu and Yamaguchi University, in a year-old piece from PsycINFO, showed that “green odor” — a 50/50 combination of 3Z-hexenol (leaf alcohol) and 2E-hexenal (leaf aldehyde) that invokes the smell of green leaves — prevents changes in blood pressure in humans caused by immersing the hand in slushy ice water. Interestingly, most of the 19 subjects’ blood pressure rose from the unpleasant “cold-pressor test,” and green odor prevented that rise; but the smell kept blood pressure steady even among the two subjects who showed an atypical fall in blood pressure from cold pressor alone. Significance here is that they were reproducing in humans an effect previously seen only in animals.

Your next thought may well be my own: What if it’s all in their heads? Oka’s department at UOEH is, in fact, the Psychosomatic Medicine Division. But we get into deep water when we talk about psychosomatic effects in animals — not that I think they don’t exist, it just gets tricky to define them. And more importantly, as long as a real physiological response begins with a smell, I think that this argument reduces to a squabble about which circuitry it follows: Does it go through the higher brain, or does it exist at a more reflexive level? Either way, clearly, it’s statistically robust enough to be real.

And aromatherapy, in my eyes, has leaped many, many notches above the now-defunct LifeGard. I’ll be damned.

[1] Ten of 10 for cool, though: The larger version looked like a Klingon disruptor.
[2] Wild, but true: if your receiver’s antenna isn’t close to an even fraction of the wavelength you want to pick up, it can’t capture the radio waves. If your transmitter’s antenna is the wrong size, the radio waves can’t get out, and will bounce around inside your radio, eventually frying the works.
[3] Disclaimer: though it’s obviously new to me, there appears to be a body of literature on this topic of which I was blissfully unaware. Apologies to the aromatherapy folks.

Wednesday, January 21, 2009

Instinct Defined?

Heather and I have — or had, anyway, in the early ’90s — a bête noire.

I think that I disliked the man instinctively — or at least had deep reservations — the instant I saw him. It may sound strange coming from me, a round-spectacled dude. But there’s a kind of passive-aggressive (or sometimes just aggressive) asshole who takes on the crunchy-earthy, round-glasses demeanor to hide a level of human empathy that would make the average Wall Street speculator seem grandmotherly.

He was a sales clerk at the main L.L. Bean store in Maine, and as I said I got a weird vibe from him the second Heather and I came into the backpack section. I was looking for a new external-frame pack to take on what was to be a life-transforming, 45-mile through-hike of Isle Royale National Park. He took me to a stand of internal-frame packs, and asked me why I was set on an external.

“Educate me,” he said. I wondered if he realized that I wasn’t dumb enough to miss his patronizing tone.

I don’t remember what I said, but it was probably something along the lines of not wanting to pay more for a pack that wouldn’t tolerate the kind of itinerant-tinker style of packing I was doing at the time. I certainly don’t feel all that strongly about the issue and, both then and now, owned more than one internal frame pack. What turned me off was the feeling that he wanted me, the customer, to justify my choice.

Needless to say, we didn’t buy. But the punch line was that every time thereafter, whenever we visited the main store, we ran into him. Didn’t matter what department; didn’t matter what day of the week, he’d be there. And his looks were nondescript — or maybe context-typical — enough that at first we wouldn’t recognize him. We’d be talking to a clerk, and at some point in an increasingly off-putting conversation we’d realize it was him.

One time, at least, I did recognize him instantly. I saw him, about halfway across the fairly large main floor, standing behind a service counter. Our eyes met, and — I swear to God — he pounded the counter with both fists, like a cave man. It might have been intimidating if it hadn’t been so damned distractingly bizarre.

All of which to make the point that I don’t know why the first encounter with this fellow started off on a weird footing — but a lot of researchers are open to the idea that smells may have something to do with it.

I’m talking about pheromones: airborne chemicals that members of the same species use to communicate with each other. The signal can serve sexual, social, alarm, or even identification purposes. Though the concept is hardly new — people have realized for thousands of years that a bitch in heat will draw males by smell — the word, and the specific definition of what a pheromone is and does, is exactly 50 years old this month, a bit of trivia I learned from an essay in Nature by Oxford pheromone expert Tristram D. Wyatt.

People use the word “instinctive” quite a bit without stopping to think of how meaningless it is. When we say an animal — let alone a human — reacts instinctively to a situation, in fact we’re saying we don’t have the slightest idea how or why she reacted that way, only that, in retrospect, the reaction was appropriate for reasons she couldn’t have thought through.

But what, then, is instinct? How does anybody’s “gut” tell them how to react, often in a split second? Subtle visual cues? ESP? Or are we just judging after the fact, putting a label on lucky guesses and conveniently ignoring the times our guts told us wrong?

The existence of pheromones in species from bacteria to mammals argues that, at least for some situations, the cue may be chemical. And it may be very real.

Now, pheromones are a tricky business in vertebrates, and especially mammals. In insects, where the concept was first defined, it couldn’t be simpler: a male moth encounters a female’s pheromone plume in the air, and he homes in to do the dirty (or try to talk her into it). Signal, response. Easy to conceptualize, easy to measure.

Problem is, mammals don’t do anything for just one reason. So rather than a clear signal/response arc, you get increased interest — subtle, and hard-to-pin-down, changes in behavior as each of the many factors that weigh in on what the animal does next gets its vote. Another level of complication comes from the fact that pheromone signals aren’t single molecules; they’re tightly controlled bouquets of different chemicals, containing strict ratios of each component. That’s one reason the same molecule can be a vital component of the come-hither signal for both Asian elephants and 140 kinds of moth, as Wyatt points out.

The big question, of course, is whether human beings respond to pheromones. We have a number of tantalizing hints that we do:

That last one deserves some elaboration. A lot of people have the common-wisdom version of this story: In 1971, Martha K. McClintock reported that women who lived together tended to synchronize their menstrual periods. In the popluar imagination, this has translated into: when women room together, they have their periods at the same time.

The key word, though, was “tended;” the kind of lock-step synchronization that people often assume doesn’t happen more often than we’d expect by chance. What McClintock actually saw was that, when women lived together, their periods would move toward each other a modest amount that was statistically greater than random chance. She didn’t actually see anything like perfect timing. And subsequent researchers have questioned whether the phenomenon even takes place.

Which is not to say that humans don’t use pheromones to transmit chemical messages that transcend verbal or visual cues. I think that we’re likely to see that idea verified by more data; but the jury’s still out at the moment.

Still, I know there was something I didn’t like about that guy ...

Wednesday, January 14, 2009

The Smell of Fear

Looking back on it, the kid was rattled — I think that, even through the distortion of hindsight, that’s a valid memory.

I was taking a break from dog tasks in the search for a missing adolescent girl. When I showed up at the command post, exhausted but willing to fly a desk for a while, they put me in the Operations Section, and set me to debriefing teams as they came in from the field.

A bit about debriefing: it is perhaps the second-most important part of any search, behind the “planning” function of deciding where to assign teams to search. Most searches that I’ve seen fail did so not because a searcher missed the subject, but because no searcher ever got within a reasonable detection distance of the subject. The second-most common mode of failure, though, is when important information fails to make its way from the field to the Planning Section.

The very young field-team leader before me had been assigned to search a neighbor’s farm. He and his grid team of human searchers had found nothing in the parts of the task he’d been able to accomplish; but what really stood out was how emphatically he pushed the need to send other searcher back to check the parts he couldn’t access because of locked gates.

“You need to check those fields,” he’d said. Yes, even then I was surprised that such a common on-task occurrence had fazed him so badly. Patronizingly, I admit, I put it down to his youth and inexperience.

Still, I dutifully entered, in the debrief section on the flip side of his TAF, or task-assignment form, something like: FTL reports parts of areas inaccessible: advises follow up.

That was enough, I know, to get the area searched again, and it was, more than once. What I hadn’t known at the time was that Heather — we’d arrived at the search separately, and were on different sleep cycles so I hadn’t seen much of her — had taken a dog out to that same farm the previous night, and had also come back rattled, reporting an inability to get to the whole farm but begging Ops to get another team in there.

Heather don’t rattle so easily.

They were both, in retrospect, absolutely correct to be agitated: as the parallel police investigation began to center on the caretaker for that farm, the incident gradually became a criminal search. Eventually, searchers found our subject’s quad rider, buried in manure, which along with a confession helped lead the police to a not-so-shallow grave that the perp had dug with a back hoe.

Now I don’t know whether odor had anything to do with the instinctive fear both team leaders apparently experienced at a murder scene with a body nearby. Certainly, there’s a literature linking odors to fear responses. And it beats the hell out of ESP as an explanation. But the story did come to mind when I encountered a surprising paper linking mad cow disease with fear responses.

Bruno Lobão-Soares and colleagues at the University of Sao Paulo were looking at the curious effect of a gene called Prp-C. This gene makes a protein that becomes, all by its lonesome, an infectious agent merely by becoming misfolded. In ways we don’t completely understand yet, the misfolded, Prp-Sc, protein gains the ability to convert its brethren to its derangement. This chain reaction causes a slow buildup of junk protein in the brain that leads to massive destruction of tissues and a variety of brain-wasting diseases: mad-cow, or bovine spongiform encephalopathy, in cattle; scrapie in sheep; chronic wasting disease in deer and moose; and Creutzfeldt–Jakob disease in humans.

This is a particularly frightening family of disorders, because unlike viral or bacterial infections, there is no living (if I can push the definition, for a virus) pathogen you can kill to prevent spread. Because the disease comes from an infectious protein, called a prion, it’s extremely difficult if not impossible to make contaminated tissues safe to eat: heat, radiation, chemical agents all tend to leave the Prp-Sc protein unscathed, and able to devastate a mind, if consumed. That’s why reports of scrapie, or mad cow, result in thousands of animals being killed, burned, and buried rather than eaten; it’s why hunters across the country are examining their kills carefully before they have them butchered.

But here’s the interesting thing the Brazilian Beat saw when they exposed PrpC-genetically engineered mice to a snake:

  • Both mice engineered to lack the protein and those engineered to have unnaturally high levels of it were singularly incautious in the presence of snakes (one of the signs being decreased sniffing behavior). But there was an important difference:
  • The over-endowed mice were hyperactive, fearless, and bizarrely motivated to seek sensual stimulation — by sight, touch, and smell.
  • The mice without the protein simply had no ability to concentrate on anything, fearful or otherwise — as the authors wrote, “... the Prp-C deficiency might lead to attention deficits.”

Now, linking prions to fear responses — which we know can be initiated by smells — would have been enough to warrant notice in these august pages. But what really floated this bloggist’s boat was the tricky similarity of symptoms brought about by completely opposite problems.

I’d be remiss if I didn’t note that I’m winging it here — I might be all wrong — but I’m fascinated by the possibility that two people we might call, at first glance, “impulse-control impaired” or even “hyperactive” could be acting that way because of profoundly different causes. One has diminished ability to attend to details; the other has an over-active urge to focus on them.

Isn’t science fun?

But I’ll tell you one thing: the next time I sense that level of anxiety in an incoming team leader, I will add to the TAF: FTL wigged out as all shit — this may be something real.

Friday, January 9, 2009

There Ain’t No Justice

I wasn’t there to see it — I think I was out on a task with my airscent dog, Pip — but better-half Heather described it vividly enough that I think I have a pretty accurate picture in my mind.

As the dog handler stood in front of her, waiting for a search task in a search for a missing turkey hunter, she reviewed the standard form that our search managers have all dog handlers fill out. It lists a number of the possible certifications that a search-and-rescue dog-and-handler team might have earned to demonstrate their competency in searching for lost people.

Too often, “dog handlers” check off none of the boxes; I think that I’ve been able to check two of them [1]; we’re happy when we get handlers with one cert.

This guy had checked off pretty much every one.

“I notice you have DCNR certification,” I can clearly hear Heather saying in my head. God, I know that tone — head for cover when you hear it, brother. “Who was your evaluator?”

“I can’t remember his name,” came the foolish reply. He should have been running, not talking.

“That’s interesting,” she said, “because I’m the only DCNR evaluator in this part of the state, and I don’t remember you.”

What happened next may have not exactly been dramatic — the search managers simply didn’t give this individual a search task to perform — but over the next few weeks the bureaucratic equivalent of a Tsunami came down on this truth-impaired individual. Posts to NASAR and DCNR confirmed he’d never tested with either organization; both sent him cease-and-desist letters. And as far as I know, nobody who’s at all integrated into our local SAR command structure has used him since.

“Heather is my hero,” Vicki Coup, the head canine evaluator for DCNR, told me later.

It was one of the few times I’ve ever seen bald-faced misconduct in the SAR community meet with anything approaching effective punishment, at least from the community (a few have fallen afoul of the courts, I’m afraid); probably that Vicki had ever seen, too. That’s because even fairly serious misbehavior tends to elicit only groans and protestations that “nothing can be done.” People just don’t have the stomach to be the bad guy, to take the time and trouble to follow through on dispensing punishment. In this case, the consequences happened only because Heather wouldn’t let it go.

But here’s the rub: in Pennsylvania’s decentralized commonwealth system, many of the local law enforcement agencies are not well connected to the SAR community and don’t know whom to turn to — and whom to avoid — when they have a lost person on their hands. For all I know, someone may be counting on this guy, who maybe doesn’t know the first thing about what he’s doing and certainly isn’t trustworthy, right at this moment.

And that, boys and girls, is the topic of my late-but-under-the-wire post today [2]: The cost of punishment to the punisher, and the difficult scientific job of understanding why people and other organisms who get no immediate payoff from doing anything but look out for Number One engage in altruistic behavior.

My spur was a recent paper by Hisashi Ohtsuki, of Tokyo Insitute of Technology and Kyushu University, and crew in Nature, and accompanying commentary by Bettina Rockenbach of the Univesrity of Erfurt, Germany, and Manfred Milinski of the Max Planck Institute for Evoltionary Biology in Plön. Ohtsuki carried out a series of theoretical calculations to see what conditions could make the cost of punishing a bad guy persist.

There’s a broad mismatch between scientific thinking on altruism — which stipulates that, in order to have evolved in the first place, altruism somehow has to offer a payback that covers (or exceeds) its cost — and common-sense observation. As a one-time biologist myself, I can’t get around the former. On the other hand, folks like Kristen Monroe have written pretty persuasively about the latter: how this theoretical underpinning sometimes falls on its face in view of how people actually act.

What the current papers point out is just how far (assuming they need to get payback) people will go to avoid taking on the cost of punishing someone. While I’m with Monroe in the larger sense, I have to admit that my own experience in the world of policing misbehavior among volunteer SAR responders tracks with Ohtsuki findings.

The problem, in the Tokyo-led group’s computer simulations, is that punishing a bad guy costs the punisher enough that rational decision makers would tend to avoid it [3]. If there’s little or no information on who the good guys and bad guys are, everybody tends to assume the worst and simply look out for themselves. If there’s reasonably good identification of good guys and bad guys — particularly when people can trade information on past behavior — the best strategy ought to be “CD,” for cooperate with good guys and defect, or just decline to work with bad guys. This is probably the most common strategy I see in real life among SAR teams faced with problem teams and individuals.

But. There is a Goldilocks set of conditions — when reasonably good information is available on past behavior, and punishing bad guys can establish you as a good guy — in which “CP,” for cooperate with good and punish bad — can arise and persist.

“Heather is my hero,” you may recall the head DCNR evaluator telling me.

Even more interesting was a bit that Rockenbach and Milinski added, to the effect that their own research suggests that CD and CP can co-exist and were actually more stable than CD-only, when participants had the option of avoiding bad guys but punishing really bad guys.

Again, for reasons that I hope to write about in the not-too-far, Monroe poses some serious challenges to the simple idea that there has to be a payoff; even on a theoretical basis, Rockenbach and Milinski say of the more complex regime in which reaction to bad guys is conditional, “Such a possibility sets a challenge for theorists.” And Ohtsuki and pals admit that a system can persist, for quite some time, even when it’s not completely optimal.

Read Heart of Altruism if you want to know more right away about Monroe’s work, I can certainly recommend it. But I’m damned if the virtual actors in these latest reports aren’t, under what amounts to “what’s in it for me” rules, acting a lot like some people I know.

One word of advice, though: if you’re going to piss somebody off, try not to make it Heather.

[1] Pennsylvania Department of Conservation and Natural Resources Air-Scent K9, and my current cert, National Association for Search and Rescue Canine SARTECH II.
[2] I’m embarking on new territory, seeing if I can get permission to republish figures from several journal articles for a series on the structure of airborne scent plumes and what nature has to tell us about how best to follow them to source. Wish me luck.
[3] Yes, I know. The trick is, then, that irrationality must offer a payback that makes it persist, which almost takes you back to square one — though it can in fact offer a payback for some other context, unrelated to altruism, that offsets its cost in the altruism context.