Wednesday 23 December 2015

“More GABA” for Autism and Epilepsy? Not so Simple

Today’s post was prompted by Tyler highlighting a very recent paper from MIT and Harvard, with some interesting research on GABA in autism.  It also provides the occasion to include an interesting epilepsy therapy, which I encountered a while back.  This fits with my suggestion that the onset of much epilepsy in autism could be prevented.

In the MIT/Harvard study, they were looking into the excitatory/ inhibitory (E/I) imbalance found in ASD and schizophrenia. They used a non-invasive optical method to measure E/I imbalance and this did get some media coverage.  However, I am not sure this could be a diagnostic tool in very young children with classic autism, as was suggested; most such children would not cooperate.  It is not just a problem of being non-verbal, as was suggested in the media.

Indeed, due to the nature of the experiment, the researchers involved older subjects, with milder autism and none had MR/ID (IQ<70).  Being a trial done in the US, of the 20 autistic subjects, 11  were being treated with psychiatric medications: antidepressants (n = 8), antipsychotics (n = 2), antiepileptics (n = 4), and anxiolytics (n = 2).

The easy to read version is from the MIT website:-

Study finds altered brain chemistry in people with autism

The full version is here:-

They used something called Binocular Rivalry  as a proxy for  E/I imbalance.

During binocular rivalry, two images, one presented to each eye, vie for perceptual dominance as neuronal populations that are selective for each eye’s input suppress each other in alternation [16, 17]. The strength of perceptual suppression during rivalry is thought to depend on the balance of inhibitory and excitatory cortical dynamics [12–15] and may serve as a non-invasive perceptual marker of the putative perturbation in inhibitory signaling thought to characterize the autistic brain.

We therefore measured the dynamics of binocular rivalry in individuals with and without a diagnosis of autism (41 individuals, 20 with autism). As predicted, individuals with autism demonstrated a slower rate of binocular rivalry (switches per trial: controls = 8.68, autism = 4.19; F(1,37) = 16.52, hp 2 = 0.311, p = 0.001; Figure 1A), which was marked by reduced periods of perceptual suppression (proportion of each trial spent viewing a dominant percept, (dominant percept durations)/(dominant + mixed percept durations): controls = 0.69; autism = 0.55; F(1,36) = 7.27, hp 2 = 0.172, p = 0.011; Figure 1B). The strength of perceptual suppression inversely predicted clinical measures of autistic symptomatology (Autism Diagnostic Observation Schedule [ADOS]: Rs = 0.39, p = 0.027; Figure 1) and showed high test-retest reliability in a control experiment (R = 0.94, p < 0.001; see Supplemental Experimental Procedures and also [18]). These results replicate our previous findings in an independent sample of autistic individuals [11] and confirm rivalry disruptions as a robust behavioral marker of autism.

To test whether altered binocular rivalry dynamics in autism are linked to the reduced action of inhibitory (g-aminobutyric acid [GABA]) or excitatory (glutamate [Glx]) neurotransmitters in the brain, we measured the concentration of these neurotransmitters in visual cortex using magnetic resonance spectroscopy (MRS).

GABA and glutamate are predicted to contribute to different aspects of binocular rivalry dynamics: mutual inhibition between (GABA) and recurrent excitation within (glutamate) populations of neurons coding for the two oscillating percepts [14].

. Critically, reducing either mutual inhibition or recurrent excitation is predicted to reduce the strength of perceptual suppression during rivalry in one implementation of this model [14], mirroring the dynamics we observed in autism. We therefore considered each neurotransmitter separately to test whether inhibitory or excitatory signaling was selectively disrupted in the autistic brain.

As predicted by models of binocular rivalry, GABA concentrations in visual cortex strongly predicted rivalry dynamics in controls, where more GABA corresponded to longer periods of perceptual suppression (Rs = 0.62, p = 0.002; Figure 2B). However, this relationship was strikingly absent in individuals with autism (Rs = 0.02, p = 0.473; Figure 2B). The difference between the two correlations was significant (hp 2 = 0.167, p = 0.013; Figure 2C), indicating a reduced impact of GABA on perceptual suppression in the autistic brain.

GABA was working backwards

Importantly, this finding was specific to GABA: glutamate strongly predicted the dynamics of binocular rivalry in autism (Rs = 0.60, p = 0.004; Figure 2B), to the same degree as that found in controls.

Glumate is working just fine.

These findings suggest that alterations in the GABAergic signaling pathway may characterize autistic neurobiology. Consistent with prior evidence from animal and post-mortem studies, such dysfunction may arise from perturbations in key components of the GABAergic pathway beyond GABA levels, such as receptors [3–9] and inhibitory neuronal density

Together with the pivotal roles of GABA in canonical cortical computations [39] and neurodevelopment [40], these findings point to the GABAergic signaling pathway as a prime suspect in the neurobiology of this pervasive developmental disorder [41]

This study reconfirms what regular readers of this blog already knew.


I thought it was positive that the MIT researchers suggested that the high level of epilepsy in autism and this E/I imbalance really must be connected.

I have been suggesting for some time that by correcting this E/I imbalance in children with autism, it is likely that the onset of epilepsy could be avoided (in some cases).

I did suggest this to one well known researcher who thought the idea of preventing the onset of epilepsy was not something that the medical community would accept as a concept.

I also raised the novel epilepsy therapy, below, to the same researcher who thought it also would never be considered.

The therapy was to use both bumetanide and potassium bromide to switch GABA back to inhibitory and then give a little boost using a GABA agonist.   

There are many types of epilepsy and some do not respond well to current treatments.  It would seem plausible that the autism-associated type of epilepsy might constitute a specific sub-type.

Potassium Bromide was the original epilepsy therapy over a hundred years ago.  It is still used in Germany as a therapy.  Reports from a century ago suggest it has the same effect in autism as Bumetanide. (we saw this in my post on autism history). 

As you can see on Wikipedia there is a wide range of GABA agonists, but the only ones that would help in epilepsy and autism would be the ones that can cross the blood brain barrier.

GABAA receptor Agonists

·         Bamaluzole
·         GABA
·         Gabamide
·         GABOB
·         Gaboxadol
·         Ibotenic acid
·         Isoguvacine
·         Isonipecotic acid
·         Muscimol
·         Phenibut
·         Picamilon
·         Progabide
·         Quisqualamine
·         SL 75102
·         Thiomuscimol

In an earlier post, we looked at the possible use of small doses of AEDs (anti-epileptic drugs).  One reader found that tiny dose of Valproate (known to raise GABA) had a positive effect when combines with Bumetanide.

In a recent comment one reader showed the same result by combing picamilon with bumetanide.

Both Picamilon and Valproate are having the effect proposed by the epilepsy researchers.

Potassium Bromide does have known side effects, but the idea of further boosting the effect of Bumetanide is interesting.  I have suggested before that this should also be possible using Diamox (Acetazolamide).  Diamox does not affect NKCC1 or EGABA,  it affects the  Cl-/HCO3-exchanger AE3  to further affect Cl- levels.  

I did suggest this a long time ago in my posts on the GABAa receptor.  I am not the only one to realize this.

NKCC1 and AE3 Appear to Accumulate Chloride in Embryonic Motoneurons


Picamilon is well researched Russian drug, sold in other countries as a supplement.  It is a modified version of GABA that includes niacin; together it can cross the blood brain barrier (BBB).

So I think a better version of what the epilepsy researchers suggest might be:-

                           Bumetanide  +  Diamox  +  a touch of Picamilon

What would be the effect in autism?

Wednesday 16 December 2015

Long Term use of Low Dose Clonazepam and More Science on the Excitatory/Inhibitory Imbalance in Schizophrenia and ASD

A small number of readers of this blog have followed Professor Catterall’s ideas and trialed low dose clonazepam for autism.  

This post summarizes my findings from using it long-term; it would be a good place to collect the findings of other people.

The science part of this blog is courtesy of a reader who highlighted the full-text version of a paper I mentioned.  Perhaps it was the author?

For information on Catterall’s clonazepam research, go to the “Index by Subject” tab and click on Clonazepam.

Before getting to that, I do get asked how I know, for sure, these therapies really do work for Monty, aged 12, with classic autism.  As I told Ben-Ari, the Bumetanide researcher, the best way to convince the doubting public will be to measure IQ, not autism.  If you can add 30 to 50 points to your IQ result, even the sceptics would pay attention.

I am not measuring IQ directly, but I do note things like spelling tests, math tests and handwriting.  The first pleasant surprise was actually reaching the point of sitting the same tests as the NT kids. Piano playing is another interesting proxy.

Monty’s one to one Assistant (and pal) from age 3 to 9 came to visit the other day and could not believe what his handwriting now looks like.  She had spent hundreds of hours with him practicing fine motor skills, like pencil control.  The end result was handwriting, but even then not like that of his peers. 

Cursive handwriting is now great.  Spelling tests and “quick-fire” math tests are also great.

As we now know, 20% of people diagnosed very young with quite severe autism seem to make wonderful progress.  This has happened by 5 or 6 years old, while the brain is still highly plastic.  Spontaneous accelerated development thereafter rarely seems to happen.  Monty started his Polypill therapy at the age of 9 years, in December 2012.

This is a spelling test from school, given to NT (neurotypical) ten year olds and 12 year old Monty (on paper without lines).  It is not rocket science and big brother could probably have got 20/20 in this test when he was eight years old.  But when Monty was eight years old, he was trying to break the windows of my car with his head and his handwriting did not look like this.

I have all the proof I need that modulating the excitatory/inhibitory imbalance in Monty’s autism is well worth the effort.  The effects are reversible if you stop the therapy, as should be the case.


Here I am repurposing an existing drug for a different use, at a dosage so low it is highly unlikely to cause side effects.  This is mirroring the use of the same drug, at similar low doses, in mouse models of autism by Professor Catterall.

Clonazepam at “high” doses is widely used already in people with autism, to treat seizures and extreme anxiety.  

Catterall showed that the drug has a totally different effect at very low doses (less than 10% of normal), via a specific mechanism which he has identified, the positive modulation of the α 2,3  subunits of GABAA receptors. 

GABAA receptors are made up of five sub-units, the strict composition does indeed vary over time, just to make things even more complicated.  The most common GABAA receptors have two αs, two βs, and one γ 2β2γ). For each subunit, many subtypes exist (α1–6, β1–3, and γ1–3). It is these subtypes of the subunits that Catterall showed to be key.  Clonazepam was one of the substances that he showed to be effective (in mice).

At “high” doses Clonazepam does have side effects, people build up tolerance to it and so take ever higher doses, and then they get hooked on it.

At very low doses the reverse seems to occur.  Over time you become more sensitive to it and need lower and lower doses.  This was a surprise to me.

The other surprise was that slightly above the effective “low dose” you get some anxiety and irritability.  When I first wrote about this I did wonder if this was just a coincidence, but it is not.

My chart from back then:-

Another interesting point was that some other readers found the effective dose was even less than mine.

When you read about the use of Clonazepam at regular, much higher doses, it is clear that there are wide variations in people’s sensitivity to this drug.  So much so that there is standard lab test to measure blood concentration of this drug, so that the clinician can vary the dose to achieve the desired level in blood.

It is not an expensive test and I did wonder if this could be used by clinicians to find the effective low dose in their patients with autism.

It did sound a clever idea, but then I read that even the same blood concentration of clonazepam (at high doses) can have markedly different effects in different people.  Still it is better than doing nothing and would reduce some of the guesswork with dosage.

The effective dose

In my n=1 example, the effective dose started out at 40mcg a day.  The half-life is very long and so you need three days to reach a stable level.

Other people contacted me to say that in their case 25mcg a day was effective and in one case, dosage once every two days was optimal.

In my case 40mcg, now gives the negative effects I has originally discovered at higher doses.

Currently the effective dose is 20 to 25 mcg.

This is a tiny dose, technically sub-clinical, but it really is better than giving none.  I have discontinued on several occasions.  There is cognitive loss, which is then regained when re-starting. 

The incremental cognitive effect is not as great in magnitude as I found with Bumetanide, but in people not using Bumetanide, the effect seems to be much greater.  Put more simply, Clonazepam plus Bumetanide is more beneficial than Bumetanide alone, at least in my case.

At this dose the annual cost of the therapy is one dollar/euro/pound. So it will not break the bank.

Tablets are available as 0.5mg  (giving 20 days of use) and 2mg (giving 80 days of use).  A bottle of 2mg tablets will last someone a few years.

I wish they made 0.025 mg (25 mcg) tablets.

I see no reason why, in ten to twenty years’ time, low dose clonazepam will not be a mainstream therapy for some autism; the only problem is the variability of the effective dosage.


For those diehards who have made it this far, now I move from the Peter-reviewed science to the Peer-reviewed science, but from yet another Peter, Peter Penzes from Northwestern University, close by the Windy City.

Abstract: Autism Spectrum Disorders (ASD) and Schizophrenia (SCZ) are cognitive disorders with complex genetic architectures but overlapping behavioral phenotypes, which suggests common pathway perturbations. Multiple lines of evidence implicate imbalances in excitatory and inhibitory activity (E/I imbalance) as a shared pathophysiological mechanism.
Thus, understanding the molecular underpinnings of E/I imbalance may provide essential insight into the etiology of these disorders and may uncover novel targets for future drug discovery. Here, we review key genetic, physiological, neuropathological, functional, and pathway studies that suggest alterations to excitatory/inhibitory circuits are keys to ASD and SCZ pathogenesis.

This study really shows how the common genetic dysfunctions in both schizophrenia and autism come together to produce the Excitatory/Inhibitory (E/I) imbalance.  Numerous different dysfunctions result in the same imbalance, some relate to GABA and some to NMDAR, but the end result is the same.

It is a really good paper, mentioning many of the genes we have encountered in this blog, plus many of the pathways like mTOR and even PAK inhibitors.

The study does not cover any therapeutic methods to correct the E/I imbalance, but this blog has those in spades.  They relate to modulating GABAA, GABAB and NMDA receptors.

Low dose clonazepam is modulating GABAA , as does Bumetanide and as should Acetazolamide (Diamox).  More of that in 2016.

Friday 11 December 2015

Treatable ID and Some Autism

Vancouver is one of the most attractive cities I have visited.  It is home to BC Children’s Hospital and Dr Sylvia Stöckler-Ipsiroglu and Dr Clara van Karnebeek. Together they have produced a remarkably thorough website called Treatable-ID, which sets out information on 82 treatable forms of Intellectual Disability (ID), formerly known as Mental Retardation (MR).
This excellent resource was recently brought to my attention by a reader of this blog from Down Under, another place well worth visiting.  Thanks, Alexandria.

ID/MR and Autism

ID/MR is defined as having an IQ less than 70; this means the cognitively weakest 2.2% of the population.

Classic Autism, Autistic Disorder or what we might now also call Strict Definition Autism affects about 0.3% of the population.  It is likely that about half of this group would score <70 in an IQ test.  I do not suggest they take one.

It is clear that an overlap might exist between the causes of MR/ID and the cause of some Strict Definition Autism.

In earlier posts I have referred to improving cognitive function in autism using Bumetanide.  We even saw that it should also improve cognitive function in Down Syndrome.

I suggested that Diamox/ Acetazolamide, another diuretic, could also have a similar effect (via the AE3 cotransporter).  One reader of this blog, Agnieszka, has been sharing her use of Acetazolamide, in the comments on the previous post.

People with RASopathies often have autism and MR/ID.  There are potential RAS therapies, one of which is a cheap statin drug.

We saw how dendritic spine morphology could be modulated and how that could affect cognitive function.  PAK inhibitors can, in theory, achieve this.

So I am already sold on the idea of some cognitive dysfunction being treatable, but I thought I was in a minority of a few dozen. Apparently not.

A friend recently highlighted my suggested autism therapies to a leading Spanish Neurologist, who clearly thinks I am just dreaming.  What would he make of Sylvia and Clara, the BC Duo?  Too much medicinal marijuana, perhaps? 

Science is all about remaining open-minded.  This should also be true for Medicine, but very often it is not. Combine this with the reality that kids with ID/MR/autism are bottom of the list of most people's priorities and you will see why things do not change, unless YOU make the changes, for your n=1 at home. 

81 inborn errors of metabolism related to Intellectual Disability and amenable to therapy

The BC Duo have collated the data on 81 treatable forms of ID/MR.

Not surprisingly some of these 81 also lead to “autism”, so they must also be treatable.  Roger, one this blog’s followers, has at least one of these 81.

So I suggest that anyone interested in a type of autism with some degree of cognitive impairment takes a good look at their site.

These are the 81 inborn errors:-

Not to confuse Sylvia and Clara with the other dynamic duo, your kids may know, from DC, rather than BC.

With so many treatable forms of MR/ID/Autism out there, is it not a little strange that thorough metabolic testing and Whole Exome Sequencing (WES) are not standard procedures after diagnosis?

By the way, WES is only as good as its interpretation.  Even world leading centres can be very weak in this respect. Insist on receiving the extended report and check all the possibly dysfunctional genes yourself.  It is not so hard.

Monday 7 December 2015

One of Thousands Autism

Some occasional visitors of this blog ask why if one drug helps their case of autism, another can be ineffective.

Perhaps it would be much more helpful right at the start to diagnose people with “one of thousands autism”, then people might better understand their situation,  and plan their way forward.

Autism is not a biological diagnosis, it is just an observational diagnosis.

Some readers have suggested “sorry, we don’t know” would be the honest diagnosis.

Several hundred autism genes and still counting

Some of these 740 genes linked to autism are shown here:-

There are existing mouse models covering 192 individual genes that can cause “autism”.

There are 18 known individual chemicals that can induce ‘autism” in mice

You may wonder how come there is a thimerosal-induced mouse model, but it exists.

There are dozens of rescue models, where scientists can make a mouse have “autism” and then reverse it.

Autism as an Adaptive Response

Any combination of what will likely become at least a thousand genetic and environmental factors can lead to what gets termed “autism”.  Autism is just the result of the brain’s adaptive response to those factors.

Certainly there are common pathways and downstream nexuses, where unrelated dysfunctions converge.

So in the end there will be a manageable number of clusters where most people’s autism can be located.

The thousand autisms will not require a thousand different therapies.

Sooner or later it will be necessary to stop calling it autism and start diagnosing each person’s biological dysfunction.  Only then can you treat it properly.

The efforts of the late Lorna Wing (autism as a “spectrum disorder”, ASD) and Barons Cohen (autism is not a disorder, it’s a friendly “autistic spectrum condition”, ASC) have certainly served to dramatically widen the number of people diagnosed, and even now wanting to be diagnosed, but have made actually treating it, very much harder.

We should start with what Knut Wittkowski, from the previous post, called Strict Definition Autism.  Apply what doctors call triage, start with those where you can make the biggest impact;  minimize cognitive dysfunction (mental retardation), self-injury, aggression and epilepsy.

Your one in thousands Autism

If you took a representative sample of 2015 diagnosed children with “autism”, who would it contain?

The following is based on what appears in previous posts and is not supposed to be definitive.

·        The largest category is probably misdiagnosed autism 

This would include people who are naturally late at developing speech, some people better diagnosed as ADHD, some people with Mental Retardation / Intellectual Disability, people who were deaf during key developmental periods and even people brought up in a cold, stimulus free environment like a 1980/90s orphanage in Romania.  Recall Leo Kanner's refrigerator mother ideas.

Quasi-autistic patterns following severe early global privation. English and Romanian Adoptees (ERA)Study Team.

·        Metabolic disorders that are likely not caused by a single gene

The big one here is mitochondrial disease, often triggered by some environmental event, like oxidative stress or an inflammatory response.  This includes the group that had a viral infection and then regress into autism, usually before the age of five.

There are other metabolic disorders like, cerebral folate deficiency, that are substantially reversible.

·        Then comes the large number of single gene dysfunctions that lead to symptoms that often include autistic behaviors

These vary from reversible to treatable.  Some well-known and not so well known, examples are:-

·        Smith–Lemli–Opitz syndrome (also SLOS, or 7-dehydrocholesterol reductase deficiency) which is in effect low cholesterol

·        Biotinidase deficiency

·        X-linked creatine deficiency

·        Pitt Hopkins

·        Rett Syndrome

·        Fragile X

All the above can be identified by genetic testing, but often are not.

·        Then comes “Dysmaturational Syndrome” which is Tourette’s Syndrome with autism “recovery” by 6 years old

This group accounted for about 5% of diagnoses in a large Italian study.   They do maintain their tics, but the autism features just fade away.

Now we are heading to what the “experts” call “Idiopathic autism”, which is the “we really don’t know”, catch-all category.

This group I will split into hypo/hyperactive pro-growth signaling pathways, based on the clever recent suggestion of Subramanian et al, from Johns Hopkins, we saw in an earlier post.

·        Hyperactive pro-growth signaling pathways

This group includes the textbook classic autism, with accelerated (brain) growth.  They can be identified by some of the following:-

·        Noticeably big head (and brain), maybe just at birth and maybe even Chiari 1 brain hernia (caused by no space for the growing brain)

·        High birth weight and muscular tone as a one year year old

·        Subsequent large drop down the percentiles on growth charts.

·        Hypoactive pro-growth signaling pathways

This group is smaller than the Hyperactive pro-growth category, but is sufficiently large to make most autism clinical trial pretty useless.

In many ways this hypo group are the entire opposite the hyper group, and what is good for them, may well be the opposite of what is good for the others.

This group will have small brains and I presume will have low birth weight. 

This does not mean they will be small as adults.

The hypo/hyper active growth refers to what is happening as the fetus develops and in the first year or two after birth. 

Implications for Clinical Trials and Therapies

It really is not good enough to carry out clinical trials on people, based solely on a DSM behavioral diagnosis of autism.  They are almost doomed to fail and indeed they almost always have failed.

Identify sub-types of autism, based on biological markers and then make trials on one sub-type vs another sub-type.  Then we might actually learn much more.

Some interventions that work in one sub-group should actually aggravate the autism of other sub-groups.  This should be entirely expected.

Autism researchers need to wake up, read other people's research and properly plan their trials based on the entirety of what we already know.  It is not rocket science; that is actually far more complex.  Planning trips to Mars is far more complex than what most autism researchers get up to. 

If you are going to compare therapies with other autism Mums/Moms and Dads, first check that your sub-type of autism is vaguely similar to their sub-type.  Otherwise you may be wasting your time/money and possibly doing more harm than good.


After receiving a diagnosis of autism, ASD or PDD-NOS, I suggest you ask the specialist to be more specific and help find you a biological diagnosis, rather than the observational/behavioral one.

If they cannot give you, at least pointers towards, a biological diagnosis, perhaps they should not be diagnosing  autism? Or just admit “I do not read the autism literature and so I know little more than you; but I get well paid as an autism expert, regardless”.