AT: Your work has broken away from the traditional view of the condition as a genetic, brain-based disorder to look at autism as a systematic or body disorder with genetic influences that affect the brain. Yet, as you just mentioned, there are other factors, such as immune system dysfunction, chronic infection, gastrointestinal issues, and detoxification problems that, according to your research, are essential to understanding autism as a chronic disease state.

Dr Herbert: Yes. You can quote papers whose numbers range all over the place on how many children with autism have gastrointes-tinal dysfunction, just to start with that organ system as an example. It ranges from 9% to 95% or more, and it really depends a lot on how you ask the question. If you retrospectively review the records of psychiatrists who interview children with autism, you will get a very low rate, but then how much attention do most psychiatrists pay to taking a GI history? Unless they've done functional medi-cine, they're not going to do that. Other papers have figures as high as 95% or even 100%. If you prospectively look for GI problems in children with autism, deliberately seeking out signs and symptoms, you get much higher numbers. Work is starting to happen - in the Autism Treatment Network, for one - to investigate every autistic child who enters a clinic to see if they have GI problems; it will be most interesting to see what is found. These studies will have to be limited in how exhaustively they can study children who may have ambiguous or absent GI-related findings. There are ethical issues around doing invasive GI procedures in children who have no signs or symptoms. You can do a capsular endoscopy, where the child swallows a little camera that takes pictures and that's probably the least invasive thing, but really we're nowhere near being able to do that on every kid who walks in the clinic door.

I think there's an insidious and difficult to document interaction between immune dysfunctions and infection. Regarding the immune problems, an off-the-cuff estimate I got from some researchers at the MIND Institute at UC Davis is that about 70% of the autism spectrum children they characterize have immune abnormalities. Regarding infection issues, this is tough in a clinical setting. How do you test for this? The tests available in academic labs won't clarify this issue. You can walk around with a chronic infection that doesn't meet the standard lab criteria for a fulminant, bad infection, but it can still be a chronic problem that throws the system's biochemistry and immune activity out of whack. Many people who think biomedically about autism suspect that something like this is common on the autism spectrum. But this will fly under the radar screen of standard laboratory measures.

One of the things I've gotten into sociologically and philosophically, as well as scientifically and clinically, is thinking about how our assumptions shape even something as ostensibly "objective" as laboratory measures. I started to think about, number one, how do you decide what is a normal range for a laboratory measure? And number two, what sorts of things do you measure? I got into this because like many clinicians, I have sent organic and amino acids to academic labs for metabolic workups of autistic pa-tients, and they come back with lots of individual values being high or low, but with the whole study being interpreted as normal. This frustrates even colleagues who know nothing about functional medicine. You start to be able to think about it when you add environment to your equation. If you have a metabolic pathway that is being inhibited by the environment, or an immune system that just isn't working up to par, a much lower level of noxious stimulus can get you into trouble and yet if you send a sample to a lab, they'll call it normal because it won't fit the patterns of genetic diseases that they are looking to diagnose. I think that the whole domain of environmental illness of which autism, in my view, is one example, raises these kinds of problems.

For autism, I think that the core of the paradigm shift will occur as we show concretely that metabolism relates to brain func-tion. I think this is plausible based on the kinds of psychological and brain functional issues we see. If you look at the psychologi-cal domains that are characteristics of autism, you have breakdowns in the areas that involve complex integrative information processing, a formulation Dr Nancy Minshew has done much to develop. With language, for example, you have to get so many things together at one time for proper functioning: motor production, breathing, meaning, associations, nuance. You can have a very bright autistic person who's articulate, and you can have a substantive conversation, but when you try and tell them a joke or make a sarcastic remark, often they don't get it-they can't take that extra step. They have a hard time adapting to change. This is the "repetitive and restricted behaviors and interests" part of the definition of autism. They have a hard time reading the emotion conveyed in facial expression. These are all things that require complex information processing.

One of the things that comes up here is, does this help explain the relationship between a severe condition such as autism and some of the milder conditions like specific language impairment or attention deficit disorder or other things? You see various brain features that are common between autism and developmental language disorder-same as specific language impairment-where there are language problems but where the behavioral and communication problems are absent or more subtle. You see similar problems processing sensory stimuli that come in rapid sequences. In my work, as I mentioned, we also measured larger brain and white matter volume in both conditions. Yet autism has overt challenges in more domains than the milder conditions. Is it that there's a kind of a threshold effect, a tipping point, where these disruptions that mess up the exquisite regulation of brain function start out being modest and when they cross a certain point, they start challenging and taking out more domains? Do the disruptions make it harder for you to do the things that come naturally to other people because they don't face an overwhelming physiological struggle to make the cellular machinery of integrative connections work?

In my mind, these problems have become linked to the metabolic things we are learning about autism. If you look at the brain as a functioning system, and if you have abnormal cytokines floating around your body, if you have abnormal metabolites, if you have malabsorption and you're not absorbing your zinc or a whole variety of other neurologically important substances, you're not going to be able to do brain business as usual. You just won't. It also links to environmental things. Maybe the increased numbers of reported cases of autism relate to more people getting hit hard enough by environmental overload to cross the tipping point into more complex involvement, into having their brains struggling so hard that even more domains cannot be coordinated.

And conversely, if you recognize all of this and you start fixing some of those systemic conditions, then the basic materials that will allow the brain to function will be provided again, and various of the cellular maladaptations can be corrected (either quickly or over time) and you can potentially achieve a more optimal, functioning state. Then the brain can start optimizing its system and coordinating functions. That's where you go from a systemic model to a model of treatment and potential reversibility.

It may be that the most strategic research intervention is to demonstrate in published peer-reviewed literature that these levels are yoked to each other. If brain measures that fascinate the cognitive neuroscience people can be shown to co-vary with metabolic measures that fascinate the physiologically oriented researchers, it might open the gates and increase the motivation to tackle a lot of the other issues that need to be straightened out in order to pursue these intrinsic relationships. Opening these gates is a core part of TRANSCEND's mission.

AT: Will improvements in these areas result in improvements in function for the patient?

Dr Herbert: There's a whole movement right now on autism recovery. There are people who insist that their children have recov-ered. There are claims that there is recovery through biomedical intervention, but also through behavioral intervention. I became fascinated with this concept when I met children who had substantially improved-I didn't know them before recovery, but I saw videos from when they were severely affected so either they all had unaffected identical twins or clones or they had substantially improved. I think that in the last 3 or 4 years since I first became acquainted with the notion of autism recovery I've seen a lot more receptivity to this idea. There's a lot of documentation on the Internet and increasingly in the media. But there are almost no academic papers-we are just beginning to see academic literature on recovery. Deborah Fein at the University of Connecticut is working on this; she is studying what predicts children who later lose their diagnosis and what problems are left over. For example, some kids wind up no longer autistic but are still language impaired.

The whole premise of treating autism, particularly biomedically, is that improvement and recovery are possible. The implausi-bility of these treatments to mainstream practitioners is linked to the generally framing of autism as genetically hard-wired and hope-less. Recovery may be happening, but until it appears in the peer- reviewed scientific literature, it won't be considered real. So we need to collect data and write it up. Some of my colleagues and I have organized something called the Autism Recovery Documen-tation Project, and we are intently, intensively working on collecting documentation of recovery and tabulating and analyzing it. How severe were they when they were affected? How far have they recovered? I expect we will have something drafted in the next several months.

This concept of recovery or even loss of diagnosis with milder residual has huge implications for both research and treatment. Take research. Two years ago, I was at an autism think tank of academic environmental researchers, and everybody was com-pletely convinced that autism was totally hopeless and irreversible and therefore we had the next 30 years to leisurely research it because there was nothing you could do in the short run anyway. Jill James and I talked about improvement and recovery but it went over like a lead balloon. But now, you just can't say, "assume hopelessness" in quite the same way because more people have heard something a little different. I think this has enormous implications for the research agenda. I am biased to wanting to favor hope rather than no hope. If we think autism recovery phenomena are real, then we need to validate and publish them. And we really need to be orienting research toward documenting and facilitating constructive change, and especially we need to prioritize offering and optimizing treatment now because if some kids can make it all the way or even a fair amount of the way to optimal outcome, then all children should be given the treatment that might help them to improve their functioning.

All too often when parents take their child in for diagnosis, they will hear, "Oh, it's autism. It's hopeless. There's nothing you can do; you'll eventually have to institutionalize the child." Or, "Just get some basic services and go home and don't have any ex-pectations." There are huge implications if those statements are shown to be incorrect. It means we need a major change in practice parameters. For this change to happen, the public pressure that comes from putting up websites and doing videos and films and so forth will help a lot in changing the atmosphere. But the sociology of this is that a precondition for changing practice parameters in autism to reflect the possibility of major improvement and recovery is documenting these phenomena in the peer-reviewed scientific literature.

AT: Your work has proposed a direct link between environmental toxins and their impact on common pathways as manifested in autism. But it's not a simple equation, is it, when you consider the criteria used to determine the level of baseline impairments from these toxins as well as the impact of multiple toxin exposures?