the intense world theory – a unifying theory of the neurobiology of autism

8/8/2019 The Intense World Theory a unifying theory of the neurobiology of autism

1/29

The Intense World Theory a unifying theory of theneurobiology of autism

Autism covers a wide spectrum o disorders or which there are many views, hypotheses

and theories. Here we propose a uniying theory o autism, the Intense World Theory. The

proposed neuropathology is hyper-unctioning o local neural microcircuits, best characterized

by hyper-reactivity and hyper-plasticity. Such hyper-unctional microcircuits are speculated to

become autonomous and memory trapped leading to the core cognitive consequences o

hyper-perception, hyper-attention, hyper-memory and hyper-emotionality. The theory is centered

on the neocortex and the amygdala, but could potentially be applied to all brain regions. The

severity on each axis depends on the severity o the molecular syndrome expressed in dierent

brain regions, which could uniquely shape the repertoire o symptoms o an autistic child. The

progression o the disorder is proposed to be driven by overly strong reactions to experiences

that drive the brain to a hyper-preerence and overly selective state, which becomes more

extreme with each new experience and may be particularly accelerated by emotionally chargedexperiences and trauma. This may lead to obsessively detailed inormation processing o

ragments o the world and an involuntarily and systematic decoupling o the autist rom what

becomes a painully intense world. The autistic is proposed to become trapped in a limited, but

highly secure internal world with minimal extremes and surprises. We present the key studies

that support this theory o autism, show how this theory can better explain past ndings, and

how it could resolve apparently conficting data and interpretations. The theory also makes

urther predictions rom the molecular to the behavioral levels, provides a treatment strategy

and presents its own alsiying hypothesis.

Keywords: autism, neocortex, amygdala, neural circuitry, connectivity, plasticity, NMDA, glutamate, perception, attention,

memory, emotion, valproic acid, animal model

HUMAN NEUROSCIENCE

Kamila Markram Henry Markram

IntroductIon

Frontiers in Human Neuroscience www.frontiersin.org December 2010 | Volume 4 | Article 224 | 1

Review ARticlepublished: 21 December 2010

doi: 10.3389/nhum.2010.00224

as in humans. The multi-level approach rom molecules to behavior,possible only in an animal model, allowed piecing together the di-

erent levels o the brains organization and up toward its emergent

behavior, which revealed a common and coherent theme o altera-

tions and suggested that autism could be explained as an Intense

World Syndrome (Markram et al., 2007b). Naturally, there is a vast

gap between an animal model and autism that is currently only

recognized or humans. In this paper, we thereore explored previous

studies, results, hypotheses, and theories in the light o the Intense

World hypothesis, and examined whether this hypothesis can stand

as a ormal and uniying theory o autism.

The original notion o an Intense WorldSyndrome in autism

arose, because VPA-exposed animals exhibited amplifed ear

processing and memories (Markram et al., 2005, 2008), which indi-cated that ragments o the world could easily become emotionally

aversive and be stored excessively. In strong support o this, we

ound that on the neural circuit level, VPA-exposed animals exhib-

ited enhanced neuronal reactivity and plasticity across several brain

regions, such as the amygdala and neocortex. This provided the

potential cellular and circuit explanation or how an autistic brain

could be easily trapped in a painully intense world, potentially

explaining a broad range o common autistic symptoms such as

sensory sensitivity, withdrawal, repetitive behavior, idiosyncrasies,

and even exceptional talents.

The neurobiology o autism has been researched extensively with

growing urgency and major strides and insights over the past

30 years (Rubenstein and Merzenich, 2003; Belmonte et al., 2004b;

Courchesne, 2004; Casanova, 2007; Minshew and Williams, 2007;

Amaral et al., 2008), yet no coherent neurobiologically based theory

o autism has yet emerged to explain its entire heterogeneity. A

wide range o interpretations, hypotheses, and theories has been

put orward, each casting a dierent light on an important but spe-

cifc aspect o autism. The central question is whether the spectrum

o autism is due to a spectrum o neuropathologies or whether a

single common pathology can explain the spectrum. Recently, we put

orth a bottom-up hypothesis or autism that is neurobiologically

grounded and works its way up rom the molecular, cellular, andcircuit levels toward the potential cognitive consequences, called the

Intense World Syndrome (or extensive review see Markram et al.,

2007b). The Intense World Syndrome hypothesiswas grounded in

original experiments using the valproic acid (VPA) rat model o

autism to explore possible alterations across molecular, cellular, syn-

aptic, circuit, and behavioral levels. Such experiments can only be

perormed using an animal model. This animal model was chosen

because VPA intake during pregnancy was linked to an increased

risk o giving birth to an autistic child (Moore et al., 2000; Rasalam

et al., 2005) and VPA exposure in rats has remarkably similar eects

* and

Laboratory o Neural Microcircuits, Brain Mind Institute, Ecole Polytechnique Fdrale de Lausanne, Lausanne, Switzerland

Edited by:

Silvia A. Bunge, University o Caliornia

Berkeley, USA

Reviewed by:

Matthew K. Belmonte, Cornell

University, USA; University o

Cambridge, UK

Egidio DAngelo, University o Pavia,

Italy

*Correspondence:

Kamila Markram, Laboratory o Neural

Microcircuits, Brain Mind Institute,

School o Lie Sciences, Ecole

Polytechnique Fdrale de Lausanne,Building AAB - Oce 201 - Station 15,

1015 Lausanne, Switzerland.

e-mail: [email protected]
http://www.frontiersin.org/Human_Neuroscience/10.3389/fnhum.2010.00224/abstracthttp://www.frontiersin.org/Human_Neuroscience/10.3389/fnhum.2010.00224/abstracthttp://www.frontiersin.org/human_neurosciencehttp://www.frontiersin.org/Community/WhosWhoDetails.aspx?UID=79&d=1&sname=KamilaMarkramhttp://www.frontiersin.org/Community/WhosWhoDetails.aspx?UID=74&d=1&sname=HenryMarkramhttp://www.frontiersin.org/Community/WhosWhoDetails.aspx?UID=74&d=1&sname=HenryMarkramhttp://www.frontiersin.org/Community/WhosWhoDetails.aspx?UID=79&d=1&sname=KamilaMarkramhttp://www.frontiersin.org/Human_Neuroscience/10.3389/fnhum.2010.00224/abstracthttp://www.frontiersin.org/human_neuroscience


2/29

Markram and Markram Intense World Theory

The experimentally based and common neuropathologyproposed in the Intense World Theoryis hyper-unctioning o elementary brain modules, called local neural microcircuits, whichare characterized by hyper-reactivity and hyper-plasticity, both owhich seem to be caused by a tendency or excitatory neuronsto dominate their neighbors. Such hyper-unctional microcircuits

are proposed to easily become autonomous, leading to runaway

inormation processing, over-specialization in tasks and a hyper-preerence syndrome. The proposed core cognitive consequencesare hyperperception, hyperattention, hypermemory, unctionsmediated by the neocortex, and hyperemotionality, mediated bythe hyper-unctionality o the limbic system. These our dimensions

could potentially explain the ull spectrum o symptoms in autism,depending on the severity o the microcircuit pathology in dierentbrain regions. The degree o hyper-unctionality in dierent brainregions could vary in each child depending on genetic personalitytraits, on unique epigenetic conditions, and unique sequence opostnatal experiences.

This article begins by shortly reviewing the validity o the VPArat model o autism as well as the experimental insights obtained

rom this model, beore delving deeper into an a re-examinationand re-interpretation o previous studies on human autism in thelight o these experimental results rom the animal model. Wemake the case or a unied Intense World Theoryor autism that

can potentially explain many o the varied past results and resolvemany conficting ndings and views, and by making some alsiableexperimental predictions.

VPA And Its lInk to AutIsm

Valproic acid is widely used to treat epilepsy and bipolar disorder

and is also a potent teratogen. It was rst introduced in the 60s asan anticonvulsant and later as the mood-stabilizing drug or thetreatment o bipolar disorder. Case reports started to appear in the

90s on children with Fetal Anticonvulsant Syndrome, which includedautistic traits (Christianson et al., 1994; Williams and Hersh, 1997;Williams et al., 2001). Two independent ollow-up population studies conrmed a strong link between VPA and autism with approxi

mately 10% o exposed children exhibiting ull blown autism and80% with one or more autistic eatures (Moore et al., 2000; Rasalamet al., 2005). Overall, the autism prevalence in the prenatally VPA-exposed population is approximately 11100 times higher than inthe general population assuming prevalence rates o 1091 casesper 10,000 in the general population (Fombonne, 2006; AutismSpeaks).

The physical malormations caused by VPA, such as acial

dysmorphy and ear abnormalities, indicate an early insult to the

brainstem during embryogenesis and, more specically, aroundthe time o neural tube closure. Support or the so-called brain-

stem hypothesis o autism (Stromland et al., 1994; Rodier et al.,1996, 1997; Arndt et al., 2005) originate rom a study on the brainstem-related teratogenic eects o thalidomide (Stromland et al.,1994), another prescribed drug strongly associated with autism.This study was the rst to reveal that the occurrence o autism

is strikingly high (30%) and exclusively when thalidomide intakeoccurred during gestational days 2024, which led to the conclusions that autism is associated with a brainstem injury at the timepoint o neural tube closure. This notion is supported by magnetic

resonance imaging studies which revealed brainstem hypoplasiain autism (Hashimoto et al., 1995; Ganey et al., 1988) as well asa post-mortem study on an autistic subject who exhibited severemorphological abnormalities and neuronal loss in the brain stem(Rodier et al., 1996).

While the earlybrainstem hypothesiso autism assumes that all

other brain alterations observed in autism are a consequence o

this big bang (Rodier et al., 1996) it is possible that traces o theimpact o VPA exposure could be carried by progenitors into thewhole brain (and body), and these eects would only maniest whenthese brain regions begin to develop and have to start perormingtheir unctions. Indeed, VPA is teratogenic at sub-lethal doses andenhances gene transcription induced by a variety o exogenous

and endogenous promoters by inhibiting histone deacetylase (Phielet al., 2001).

Valproic acid given continuously throughout pregnancy to rats,as or humans, has been known or some time to cause severe behavioral alterations (or review see Vorhees, 1987a,b; Wagner et al.,2006; Markram et al., 2007b). However, theses alterations couldbe conounded by the numerous other VPA-induced insults and

cognitive and motor impairments, more associated with generalized Fetal Anticonvulsant Syndromethan with autism. To target theautism component o this anticonvulsant syndrome, Rodier et al.(1996) developed a rat model or autism by specically administering VPA only during the time period o neural tube closure according to the neurological hypothesis that the brainstem is injured inautism (reviewed in Markram et al., 2007b). The administrationo a single i.p. injection o VPA (350 mg/kg) administered to pregnant dams on embryonic days (ED) 1213 results in a reduction

o the trigeminal and hypoglossal motor nuclei, loss o neuronsin the abducens nucleus and in the oculomotor nucleus (Rodieret al., 1996), which parallels losses ound in the brainstem in autism(Hashimoto et al., 1995; Rodier et al., 1996; Ganey et al., 1988).

Follow-up anatomical studies in the rat showed that VPA exposureon ED12.5 also results in a loss o cerebellar neurons (Rodier et al.,1997; Ingram et al., 2000), one o the most prominent eatures in

the autistic brain (Ritvo et al., 1986; Kemper and Bauman, 1998;Palmen et al., 2004). Abnormalities in the serotonergic system, oneo the most indicative biochemical pathological markers in autism(Lam et al., 2006), were also ound (Narita et al., 2002; Miyazakiet al., 2005; Tsujino et al., 2007). Behaviorally, prenatal exposureto VPA on ED12.5 produces the two cardinal symptoms o autismin the rat ospring: decreased social interactions and increased

repetitive behaviors (Schneider and Przewlocki, 2005; Markramet al., 2008). In the emotional domain, the ospring also exhibitenhanced anxiety (Schneider and Przewlocki, 2005; Schneider

et al., 2007; Markram et al., 2008), in the motor domain, locomotorhyperactivity (Schneider and Przewlocki, 2005), in the nociceptiondomain, lower sensitivity to pain (Schneider et al., 2001; Schneiderand Przewlocki, 2005; Markram et al., 2008), in the sensory domain,hyper-sensitivity to non-painul sensory stimulation and impaired

pre-pulse inhibition (Schneider and Przewlocki, 2005; Markramet al., 2008), and in the memory domain, enhanced eye-blink conditioning (Stanton et al., 2007) all o which are common eatureso autism described in the DSM-IV and/or in the autism literature (Sears et al., 1994; Muris et al., 1998; American PsychiatricAssociation, 2000; McAlonan et al., 2002; Perry et al., 2007).

Fonts n Hman Noscnc www.fontsn.o December 2010 | Volume 4 | Article 224 | 2
http://www.frontiersin.org/http://www.frontiersin.org/


3/29


It is oten argued that autism is a human disorder which isbased on the higher cognitive symptoms that are most commonlyassociated with autism such a theory o mind and language decitsas well as unusual human talents. Albeit in a ar more rudimentary orm, many o the high-level cognitive unctions can also be

observed and measured in much lower mammals such as rats andmice. The common thread is the neocortex, which is the source

o mammalian higher brain unctions. The microcircuitry o themammalian neocortex is remarkably similar rom mouse to manin terms o layering, types o neurons, interconnections, and long-range connectivity principles (Silberberg et al., 2002). It would bevery dicult to argue that insults and predispositions are exclusiveto human neocortex. One may also argue that such a model ignoresthe well-established heredity component o autism (or review see,

e.g., Persico and Bourgeron, 2006), but then not all homozygotetwins succumb to autism. A pure genetic argument also ignoresthe high incidence o autism reported with thalidomide and highdoses o VPA in human ospring. The most parsimonious interpretation is that autism is a poly-genetically predisposed disorderthat is triggered by an insult and that the pathology unolds dur

ing development. The low incidence o VPA-linked autism todayis another possible argument, but the doses used today (around510 mg/kg) are about 510 times lower than doses used in the

earlier times (4050 mg/kg). The animal models however usedeven higher doses (300500 mg/kg) primarily because it would notbe possible to systematically study the alterations i only 510%o the ospring are aected as reported in the earlier times. Thehigher doses seem to bring the incidence more into the 7080%range and doses above 800 mg/kg are lethal (unpublished data). It

may also turn out that these doses could be signicantly lowered icombined with animals genetically engineered with identied predisposing mutations. Nevertheless, high doses do limit the strengtho conclusions drawn rom this model and urther validation in

human autism is required.

the neurobIologIcAl bAsIs of the Intense World

theory

The Intense World Theoryis experimentally based on direct neu

ronal recordings and behavioral testing on rat ospring exposedprenatally to a single dose (500 mg/kg) o VPA on embryonic day12.5. In the course o these studies, we ocused on the neocortexand the amygdala or these reasons specied below.

The neocortex is undamental or all higher-order cognitiveunctions such as perception, attention, and memory. The entireneocortical sheet can be viewed as a collection o unctional col

umns or modules that process dierent eatures and spatial posi

tions o the sensory environment and their relationships to thebody. The remarkable property o these columns is that they arevery similar rom mouse to man and across all neocortical regionswith a stereotypical template design and only subtle variations ordierent neocortical regions and species (Silberberg et al., 2002).These columns are thereore designed to be general purpose processors and they are interlinked via short and long-range connec

tions to orm brain areas, regions, and the neocortex as a whole.These columns react to input, and their activity must be careullycoordinated across the entire neocortex to orchestrate coherenthigher brain unction. We thereore examined the alterations in the

somatosensory cortex and the medial prerontal cortex (mPFC).The knowledge o the normal somatosensory circuit in the rat is themost extensive and somatosensory abnormalities, such as increasedsensitivity to touch, are common in autism. The prerontal cortexhas received extensive attention in autism research due to its pivotalrole in executive unction, language, social cognition, and regula

tion o emotional behavior (Struss and Knight, 2002). Based on

non-invasive imaging studies during task perormance in autisticand control subjects, some studies initially suggested that the pre-rontal cortex is not suciently activated in autism (Happ et al.,1996; Baron-Cohen et al., 1999; Ring et al., 1999; Castelli et al.,2002), but more recent studies show hyper-activation in this brainarea (Gomot et al., 2008; Knaus et al., 2008; Dichter et al., 2009;Belmonte et al., 2010). As described urther below, direct measure

ments rom somatosensory and prerontal neurons o VPA-treatedospring indeed suggest that these brain regions may be hyperreactive and hyper-plastic (Rinaldi et al., 2007, 2008a,b).

The amygdala is a key part o the emotional and social brain circuits and has many unctional roles such as detecting and interpreting signs o emotional and social signicance in the environment,

modulating memory storage across multiple brain sites, establishing ear memories, anxiety, and the regulation o autonomic andhormonal responses (reviewed in Davis and Whalen, 2001; LeDoux,2003; Zald, 2003; McGaugh, 2004; Adolphs, 2006). Dysunctiono the amygdala has been related to disorders o ear processing,anxiety, and social behaviors (reviewed in Cottraux, 2005; Damsaet al., 2005; Hajek et al., 2005; Shayegan and Stahl, 2005; Blair et al.,2006). In autism research, the amygdala was studied primarily dueto its role in the processing and interpretation o socio-emotionalcues and its infuence on social behaviors (Baron-Cohen et al., 2000;

Sweeten et al., 2002; Amaral et al., 2003; Schultz, 2005; Bachevalierand Loveland, 2006; Schulkin, 2007). The very rst animal model oautism was based on lesioning the amygdala and studying the eects

on social behavior and hierarchy (Bachevalier, 1994), implying thatthe lack o amygdala activity may explain the lack o social interactions or social intelligence in autism. This view dominated the

research perormed on the role o the amygdala in autism. Parallelswere drawn between amygdala lesioned patients and autistic sub

jects (Adolphs et al., 2001), unctional magnet resonance imaging(MRI) studies revealing an insuciently activating amygdala inautistic subjects were associated with decits in interpreting otherpeoples state o minds and eelings (Baron-Cohen et al., 1999;Critchley et al., 2000; Pierce et al., 2001). However, the oppositecould also be true and lead to similar symptoms: rather than being

hypo-active or not suciently responding, the amygdala could beoverly reactive in autism. Consequently, autistic people could be

processing too much emotionally relevant inormation, includingenhanced ear and anxiety processing. The outcome could be asimilar one to a not suciently active amygdala: withdrawal anddecreased social interaction due to an enhanced stress-responseand socio-emotional overfow. Indeed, as described below our

studies on VPA-treated rat ospring indicate that the amygdalais hyper-reactive, hyper-plastic, and generates enhanced anxietyand ear processing (Markram et al., 2008). In accordance withthis, more recent MRI studies as well reveal amygdaloid hyper-activation in autism (Dalton et al., 2005; Kleinhans et al., 2009;Monk et al., 2010).



4/29


hyPer-reActIVIty

Autistic children can be overtly sensitive to sensory stimulation,including light, sounds, and touch. VPA-exposed rat ospringexhibit impaired habituation to sensory stimulation as measured bythe level o pre-pulse inhibition in vivo (Schneider and Przewlocki,2005; Markram et al., 2008). To test neuronal reactivity to network stimulation in vitro single neuron reactivity was recorded

in the somatosensory cortex, the prerontal cortex, and the lateralamygdala, while simultaneously stimulating the brain slices witha multi-electrode array.

In both neocortical areas neuronal reactivity to network stimulation was nearly twice as strong in VPA slices than control slices,which could be observed across dierent neocortical layers (Rinaldiet al., 2008a,b). Neuronal responses in the VPA-exposed amygdalawere also greatly amplied to network stimulation and in addi

tion increased and prolonged episodes o bursting behavior wereobserved, which were greater in number, requency, and duration(Markram et al., 2008). Thus, the slightest network stimulationcompared to controls, triggers a run-away-like response in theamygdala in this animal model o autism.

In order to account or this massive increase in neuronal reactivity to network stimulation several possibilities were tested. To

check i enhanced excitability o the individual neurons withinthe microcircuit could account or the hyper-reactivity to network stimulation, excitatory synaptic currents were studied inpaired neuronal recordings and revealed that the -amino-3hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) mediatedsynaptic responses o connections between neurons were actually

weaker in the VPA-treated rat ospring, which was accountedor by ewer synapses deployed (around ve synapses are normally used in these connections) in each synaptic connection(Rinaldi et al., 2008b). In addition, examination o the passiveand active conductance properties o the excitatory pyramidal

neurons revealed that pyramidal neurons required more currentto drive the neuron to spiking threshold and that the number ospikes generated or a series o current injections was lower thanin controls (Rinaldi et al., 2008b). Thus overall, when stimulatedindividually VPA-neurons were hypo-excitable, which could not

account or hyper-reactivity to network stimulation. We thentested i reduced inhibition could account or the microcircuithyper-reactivity. Inhibitory currents were increased proportionally to the increased excitation in the neocortex, indicatingthat the excitation was able to recruit a constant matching levelo inhibition without an imbalance developing (Rinaldi et al.,2008b). In the amygdala however, we ound that inhibition was

greatly reduced (Markram et al., 2008), resulting in an excitation/

inhibition imbalance. Thus, these results paint an ambiguouspicture or the hypothesized imbalance o excitation and inhibition in autism (Casanova et al., 2003; Rubenstein and Merzenich,2003). As ar as the neocortex is concerned, excitation is balanced perectly well with inhibition in VPA-treated ospring,while in the amygdala an imbalance toward increased reactivity

due to a loss o inhibition can be observed. It remains to beexamined i the loss o inhibition is the primary mechanism oproducing hyper-reactive microcircuits in brain areas that havea high incidence o inhibitory neurons, such as the amygdala orthe cerebellum.

We then examined i changes in morphology could accountor the observed hyper-reactivity. Morphological examination o3D reconstructions o somatosensory pyramidal neurons did notshow any signicant dierences in the extent o axonal or dendriticarbors, in the spine or bouton densities, and in the size o pyramidalneuron somata. There was also no change in the number o pyrami

dal neurons or in the level o apoptosis (Rinaldi et al., 2008b).

In summary, the neuronal network response in VPA-treatedospring is greatly amplied, resulting in hyper-reactivity tonetwork stimulation as a common denominator between severaldistinct brain regions (Figure 1). Several potential mechanismscan be excluded to underlie this hyper-reactivity, such as moreexcitable neurons or stronger excitatory synaptic connections, or

changes in morphology. Quite to the contrary,VPA-neurons seemto be hypo-excitable, as i they were to compensate the strongnetwork hyper-reactivity. A loss o inhibition can be a potentialmechanism to account or hyper-reactive microcircuits in some,but not in all, brain regions. In the neocortex, inhibition matchedexcitation levels.

Ater excluding all above parameters, we examined synaptic con

nectivity patterns between neurons as a potential mechanism orthe observed hyper-reactivity.

hyPer-connectIVIty

Alterations in synaptic connectivity in autism were proposed previously (Belmonte et al., 2004a; Just et al., 2004; Courchesne andPierce, 2005b). Several MRI studies suggested that long-range connections between dierent brain areas are underdeveloped in autism(Horwitz et al., 1988; Castelli et al., 2002; Welchew et al., 2005; Justet al., 2007), and by extrapolation, that short-range connections

may be overly developed (Casanova, 2004; Courchesne and Pierce,2005a,b; Courchesne et al., 2005; Mottron et al., 2006). While in thehuman connectivity can be deduced only indirectly (e.g., through

synchronization states between dierent brain regions), the animal model poses the advantage o a direct and quantitative assessment o neuronal connectivity. We examined the number o direct

connections established between excitatory pyramidal neuronsas well as between excitatory pyramidal neurons onto inhibitoryinterneurons within the microcircuitry o the somatosensory andprerontal cortex using paired neuron recordings in brain slices oVPA treated and control ospring. We ound an increase o around50% o both neuronal connection types, excitatory, and inhibitory,

in VPA-treated ospring (Rinaldi et al., 2008a,b; Silva et al., 2009;Figure 1). When examining the circuit in an unperturbed baselinestate, this hyper-connectivity was only ound or very close neigh

boring neurons conned to the typical dimensions o a neocortical

minicolumn (less than 50 m intersomatic distances). However, wealso applied an advanced over-night stimulation protocol (or moredetails see the ollowing section on hyper-plasticity) and ound thathyper-connectivity could also emerge beyond the mini-columnarrange as a result o enhanced micro-circuit plasticity in VPA-treated

ospring (Silva et al., 2009; Figure 3).In summary, VPA-treated neuronal microcircuits exhibited a

50% increase in connectivity between neurons. Hyper-connectivitybetween neurons and thus enhanced inormation fow could bethe on the underlying mechanisms causing hyper-reactivity acrossdierent brain regions.



5/29


Figure 1 | Hyp-actvty and hyp-connctvty n VPA-tatd offspn.Depicted are schematic neural microcircuits rom control (top) and VPA-treated

ospring (bottom). Brain slices were electrically stimulated through multiple

stimulation electrodes underneath the slice a multi-electrode array.Theresponsiveness to this network stimulation was recorded rom individuallypatched cells. In comparison to controls, neurons rom VPA-treated ospring

were excessively reacting to the stimulation as depicted in the exemplary

voltage traces. Further examination by recording rom pairs o neurons revealed

that this hyper-reactivity was due to the excessive connectivity in VPA-treated

microcircuits (schematically depicted by the red hal-circles).The connectionprobability was increased by approximately 50% in microcircuits romVPA-treated ospring.

Stimulating the normal network

Stimulating the VPA-treated network

Normal network reactivity

Enhanced network reactivity

Inhibitory neuronPyramidal neuron Connection between 2 neurons Stimulating electrode Recording electrode

hyPer-PlAstIcIty

Memory processes are oten altered in autism,but research in this areahas produced quite controversial results with hypothesis postulating

amnesia (Boucher and Warrington, 1976) normal (e.g., Renner et al.,

2000; Toichi and Kamio, 2002) and enhanced memory unctioning(Sears et al., 1994; Caron et al., 2004). Long-term potentiation (LTP)is the neuronal mechanism widely assumed to underlie memory

ormation (e.g.,Muller et al., 2002). We thereore examined whethersynaptic plasticity was aected in the VPA-treated somatosensorycortex, prerontal cortex, and amygdala ollowing a Hebbian pair

ing stimulation protocol. In all three brain regions, the amount oLTP was doubled in VPA-treated ospring as compared to controls(Rinaldi et al., 2007, 2008a; Markram et al., 2008; Figure 2). In theneocortex at the age o 2 weeks, LTP usually takes on a presynaptic



6/29


Figure 2 | Cllla hyp-plastcty n VPA-tatd offspn. Long-term

potentiation (LTP) is increased in neurons rom VPA-treated ospring as

compared to controls.The graphs depict the percentual increase in response

amplitude recorded rom neurons beore and ater a strong electrical stimulation.

Increases are at least two to our times higher in VPA-treated ospring than

controls across all three recorded brain regions, the somatosensory cortex (S1),

the medial prerontal cortex (mPFC) and the lateral amygdala. Results

rearranged rom original research articles (Rinaldi et al., 2007, 2008a; Markram

et al., 2008), with permission rom PNAS, Neurospychopharmacology, and

Frontiers in Neural Circuits.

100

150

50

0

S1 mPFC Amygdalan=10

500n=14 140

n=15

n=9

VPA

n=15

Control VPA

120400

10020

0 0

Control VPA Control

Incre

ase(%)

Incre

ase(%)

Incre

ase(%)

100

80

60

300

200n=12

40

orm by increasing the release o glutamate (Markram and Tsodyks,1996). This type o presynaptic LTP was normal in the VPA-exposed

animals. However, postsynaptic LTP, normally not present at thisage in the neocortex, contributed signicantly to the enhancedresponses. This additional and boosted orm o postsynaptic LTPwas ound in both neocortical layers 2/3 and layer 5 pyramidal neurons. These results indicate that glutamatergic synapses are remarkably hyper-plastic in this animal model o autism.

Previously, we ound that connectivity between neurons can alsochange as a result o each stimulation to recongure the network termed microcircuit plasticity,which is a orm o wiring plasticity(Le Be and Markram, 2006). To test the rewiring capacity o themicro-networks, brain slices rom VPA-treated and control rats

were perused with glutamate overnight and neuronal connectivitypatterns were established beore and ater the stimulation. Indeed,in brain slices rom VPA-exposed rat ospring we ound a striking increase in the rate at which neurons connect and disconnectto rewire the circuit (Figure 3). Interestingly, this increased rateo rewiring was particularly pronounced above the intersomaticdistance o 50 m, which lies beyond the mini-columnar range.Within the mini-columnar range (intersomatic distances below50 m) the VPA-treated circuit seemed to be saturated, since connectivity is already increased at baseline levels. These results indicate a remarkably increased capacity or rewiring microcircuitsas a result to stimulation and learning experiences. This suggests

that the microcircuits in this animal model o autism do not only

react excessively and modiy their existing synapses in a strongermanner, but that overall the microcircuits are also more adaptiveto stimulation (Silva et al., 2009).

hyPer-nmdA recePtor exPressIon

The glutamatergic neurotransmitter and receptor systems, particularlyN-methyl-d-aspartate (NMDA), mediates synaptic plasticity

(Nicoll and Malenka, 1999) and alterations in this system couldcontribute to the above observed hyper-plasticity. As comparedto control, the VPA-treated neocortex did not exhibit any alterations in the AMPA receptor subunits GluR1, GluR2, and GluR3

and the obligatory subunit o the NMDA receptor, NR1. However,the NMDA receptor subunits NR2A and NR2B were more than

twoold over-expressed (Rinaldi et al., 2007). Electrophysiologicalexperiments that isolated the NMDA component o the synapticcurrents conrmed a boosted NMDA receptor mediated currentat these glutamatergic synapses. We urther examined the varioussecond-messenger systems and ound an enhancement in CaMKIIexpression, which is known to mediate NMDA receptor plasticity

induction (Silva et al., 1992; Liao et al., 1995; Giese et al., 1998;Lisman et al., 2002). We did not nd any evidence at this age that

enhanced NMDA receptor levels might render neurons more vulnerable to neurotoxicity, but have not examined whether this couldcause damage later on in lie.

hyPer-leArnIng

I neocortical columns are hyper-reactive and hyper-plastic, therecould be signicant consequences or perception, attention, as wellas or learning and memory. We thereore examined whether these

changes had any impact on learning and memory tasks that dependon the neocortex. The rat whiskers are comparable to the humanngertips and each whisker is neatly represented in the barrel cortex (Woolsey and Van der Loos, 1970), a part o the somatosensorycortex. Rats use their whiskers to build spatial representations otheir environment, locate objects, and perorm ne-grain texture

discriminations (Petersen, 2007). Rats can thereore be trainedwith a reward to discriminate between a wide and narrow aperture

using their whiskers, a task which depends on the barrel cortex(Krupa et al., 2001). VPA-exposed ospring learned and memorized better to discriminate between apertures o dierent sizesthan normal control rats (Markram et al., 2007a) consistent withthe hyper-unctionality.

hyPer-feAr

It is widely established that the amygdala mediates the ormation

and probably also storage o ear memories (LeDoux, 2003) andenhances memory ormation throughout other brain regions byacting as an emotional amplier (Cahill and McGaugh, 1996). Since



7/29


Figure 3 | Mcocct hyp-plastcty n VPA-tatd offspn. Depicted are

schematic neural microcircuits rom control (top) and VPA-treated ospring

(bottom). In this experiment, brain slices were perused or 12 h with a glutamate

solution in order to stimulate the circuits and induce rewiring o connections (red

and blue hal-circles).The connectivity probability between neurons was accessedbeore (let panels) and ater (right panels) the glutamate stimulation. In controls,

the connection probability increased signifcantly within a range o less than50 m (red hal-circles), which is the mini-columnar range, but did not change in

ranges higher than 50 m (measured up to 200 m, blue hal-circles), which is a

columnar range. In VPA-treated ospring, the connections probability was already

increased within the short mini-columnar range (below 50 m, red hal-circles)

beore the glutamate stimulation and did not increase any urther ater the

stimulation (right panel), because the connection capacity was already boosted

and probably saturated to a maximum at baseline levels. Indeed, controls only

reached a similarly high connectivity probability within the mini-columnar range as

VPA-treated ospring already exhibited at baseline levels ater the 12 h stimulation

was applied. However, inVPA-treated ospring, the connectivity probabilityincreased signifcantly at ranges above 50 m (right panel, blue hal-circles),

revealing a urther remarkable rewiring capacity at the columnar range due to

stimulation a eature normal control microcircuits did not exhibit.

Control

Pyramidal neuron Connection below 50 m intersomatic distance Connection within 50-200 m intersomatic distance

Connectivity at baseline levels Connectivity after 12 h glutamate simulation

12 h glutamate simulation

VPA-treatedConnectivity at baseline levels Connectivity after 12 h glutamate simulation

12 h glutamate simulation

Glutamate

we observed enhanced plasticity in the amygdala in the VPA modelo autism, we asked whether this would also maniest in enhanced

ear memory ormation. Indeed, VPA-exposed ospring exhibitedgreatly amplied conditioned cued and contextual ear memories when tested up to 3 months ater conditioning (Figure 4).Moreover, in VPA-treated ospring, this ear generalized to another

previously non-ear evocative tone and was more resistant to extinction than in controls (Markram et al., 2005, 2008). This indicated

that VPA-treated animals store ear memories in an exaggeratedand more persistent manner, generalize learned ear more easily tosimilar stimuli and once ear to a particular stimulus congurationis acquired it is dicult to erase.

http://www.frontiersin.org/http://www.frontiersin.org/http://www.frontiersin.org/


8/29


Figure 4 | enhancd fa mmos n VPA-tatd offspn. Ospring o

VPA-treated dams exhibits normal ear conditioning during training, but

enhanced ear memories to the tone and context 1, 30, and 90 days ater

training, with the dierences becoming more pronounced with time. Results

rearranged rom original research article (Markram et al., 2008) with permission

o Neurospychopharmacology.

Tone memory

0

20

40

60

80

100

Day 3 Day 31 Day 93

Context memory

0

20

40

60

80

100

Day 2 Day 30 Day 92

Control VPA

Freezing(%)

0

20

40

60

80

100

Day 1

Training

Anxiety and phobias are known eatures o autism (Muris et al.,1998; Gillott et al., 2001; Evans et al., 2005) and were initiallyreported byKanner (1943) himsel and are urthermore widelyclaimed in anecdotal parental reports. Nevertheless, with the

exception o two studies (Bernier et al., 2005; Gaigg and Bowler,

2007), ear processing has largely been overlooked in autismresearch and this was the rst demonstration o enhanced earprocessing in an animal model o autism suggesting that this mayalso be occurring in autism. Enhanced ear memory ormationand a progressive generalization o ears could have major consequences on behavior and account or inappropriate reactions

to the environment, sudden and apparently inexplicable anxietyattacks, loss o the nesse required in social interactions, and

phobias. Over-generalization may also accelerate the progressionin autism by more rapidly limiting the repertoire o sae stimuli,environments, and situations. While decits in extinction werepreviously observed in autistic children (Mullins and Rincover,1985; Sears et al., 1994; Coldren and Halloran, 2003) and maylead to preservation tendencies observed in autism, ear extinctionwas never studied in autism. I present, a decit in extinguishing

acquired ear in autism would make it more dicult to relinquishold ears that are no longer relevant or justiable. This decitcombined with longer-lasting ear memories that are also over-generalized, could lead to a progressive and irreversible reduction

in the repertoire o acceptable stimuli and drive a complete lockdown and blanketing out o what would rapidly become a painully intense world.

the Intense World theory of AutIsm

The challenge or any unied theory o autism is to understand thecommon cause or the wide spectrum o autistic disorders andthe autistic traits that are ound in other disorders, i there is one.The Intense World Theoryproposes that autistic traits could emergei a molecular syndromeis activated that sensitizes gene expressionpathways to respond excessively to environmental stimulation. Undernormal conditions such pathways would enable enriched environ

ments to nurture brain development, but i these pathways are

sensitized, then environmental stimulation may cause exaggeratedand accelerated development o the brain in general and the glutamatergic system o neural microcircuits in particular. Microcircuitglutamatergic hyper-unctionality in the neocortex could causehyperperception, hyperattention, and hypermemory, which are

proposed as the core triad o cognitive traits common to all autisticsymptoms. Microcircuit hyper-unctionality in the limbic systemcould cause hyperemotionalityadding a orth dimension that couldscale the cognitive impact o the triad pathology. The severity on eacho these our axes could perhaps account or autism on any part o thespectrum. The sensitivity to the environment could drive the brain



9/29


to develop in a premature sequence and in a manner that enhancesunctionality.At its peak, the environment could become excessivelyintense and set in motion a systematic regression to where the brain isorced to take reuge in a highly specialized cocoonwhere extremesand surprises are actively avoided and blocked out.

The specic molecular cascade that drives hyper-unctionality inbrain microcircuits is not thought to be necessarily the same in di

erent parts o the brain since there may be dierent ways to producehyper-unctionality in dierent regions and we thereore proposea common syndrome rather than a common specic molecularcascade, or all brain regions. The characteristics o the syndrome

could be an exaggerated response to stimulation. In the brain, thiscommon molecular syndrome is proposed to have a dual eecto causing hyperreactivityand hyperplasticityo microcircuits toproduce hyperunctionality(Figure 5).

In the neocortex, the consequences could be severe, becausemicrocircuits lie within unctional modules called neocortical

columns and i these columns are hyper-unctional, the delicate

balance between intra- and inter-columnar processing would beupset. The approximately million neocortical columns in humanseach need to be precisely excited and inhibited to coordinate higherbrain unctions and complex behavior. Excessive intra-columnarprocessing, particularly during development, could enhance themost elementary sensory, motor, and cognitive processes at the

expense o being able to orchestrate symphonies o higher cog

nitive unctions. With excessive learning and memory processes,sensory regions could consolidate into overspecialized modules andlead to hyper-preerence processing. During early development,this may lead to excessive fow o inormation rom sensory areasto the higher integration areas such as the association cortices and

prerontal lobe which may cause prematurely accelerated growtho these higher order brain areas, as indeed observed in autism(Courchesne et al., 2001; Carper et al., 2002), but this is only oneout o several possible explanations or this phenomenon. Decitsin higher brain unctions such as executive control and holisticprocessing may also be better explained by overly autonomous

Figure 5 | Th hyp-fnctonal ccts n atsm.As suggested by theIntense World Theorythree etiological actors (a genetic predisposition; an

epigenetics attack in orm o a toxic insult; environmental actors during

postnatal development) cause autism by activating a molecular syndrome that

may be dierent across dierent brain regions, but that leads to hyper-unctional

microcircuits (expressed as hyper-reactivity and hyper-plasticity) in all brain

regions. Two regions known to be aected include the neocortex and amygdala

and we hypothesize that other regions may be similarly aected. The

consequences on cognitive processing include hyper-sensitivity, -perception,

-attention, -memory, -ear, and -emotionality. We propose that this leads to an

intense, ragmented, and aversive world syndrome or the autistic child, which

could account or a spectrum o behavioral abnormalities.

Molecular Syndrome

Hyper-functional microcircuits

Geneticpredisposition

t Toxic insult(epigenetic

attack)

Hyper-reactivemicro-circuits

Hyper-plasticmicro-circuits

Hyper-Sensitivity Hyper-Perception Hyper-Attention Hyper-Memory Hyper-Fear Hyper-Emotionality

Intense World Fragmented World Aversive World

Spectrum Integration Deficits Exceptional Talents Routines Withdrawalof Idiosyncrasies Repetition Social Interaction Problems


Environmentalfactors during

postnataldevelopment


10/29


elementary modules rather than decits in these areas or weaklong-range connectivity. Indeed, structural MRI indicates higherlevels o white matter in the cerebrum and cerebellum in youngchildren with autism (Courchesne et al., 2001; Carper et al., 2002),which in the light o the Intense World Theory could be interpreted as a compensatory action to coordinate excessive activity in

columns within and between dierent brain areas.

More specically, each neocortical column is known to beinvolved in processing multiple eatures o stimuli. In the visualcortex or example, such eatures include the orientation, spatialrequency, contrast or color (Hubel and Wiesel, 1962; Tootell et al.,1981). The emphasis that each column must place on processing thedierent eatures is careully crated to allow the neocortex to simultaneously specialize in processing many dierent eatures locally

and generalize in combining eatures globally. Hyper-reactivityand hyper-plasticity could act synergistically during experiencesto enhance sensitivity to eatures and consolidate memories oeatures processed. Selective eature sensitivity would more easilyallow specic eatures to trigger column processing, but could makeit more dicult to activate other eatures (and tasks) and to inter

rupt any processing, once started. Overly strong consolidation omemories o the processing o eatures in development could alsorapidly lead to dominance o the earliest eatures (impressions)and avoidance o processing o other eatures (maniesting in higherthresholds). Such hyper-autonomous and overly selective columnscould make bottom-up control o the activation o columns rom

the thalamus and top-down control rom higher associational areasmore dicult leaving the neocortex ragmented into independentmodules that are dicult to control and coordinate an exaggerated and runaway response to stimulation.

Hyper-preerence processing in the sensory domain could leadto exaggerated selectivity, sensitivity, and specialization o sensory eatures and hence hyperperception, while hyper-preerence

processing in cognitive processing in general could lead to hyperattention and in the memory domain, to hypermemory. Thereore,the degree o hyper-preerence processing in neocortical regions,areas and columns and the normal variation o individuals, couldcontribute a spectrum o autistic traits all maniesting as part oan Intense World Syndrome.

The molecular syndrome, however, also seems to extend beyondthe neocortex.While in the neocortex this drives hyper-unctionality by hyper-connecting neighboring neurons to produce excessiveexcitation and by hyper-expressing NMDA receptors to produceexcessive plasticity, in the amygdala inhibition is also reduced.The amygdala contains relatively higher numbers o inhibitory

interneurons than in the neocortex and reducing inhibition may

be more eective in causing hyper-unctionality than hyper-connecting through glutamatergic synapses. Hyper-unctionalityin the amygdala (and related emotion centers) could add a veryimportant hyperemotionaldimension to the triad pathology torender the already intense world progressively more painul and

aversive with each experience leading to a progressive lock downand social withdrawal.

While the Intense World Theory is primarily based on experimental data derived rom the neocortex and the amygdala, we donot exclude that the same molecular syndrome driving hyper-reactivity and hyper-plasticity could be active in other structures

o the orebrain or the mid- and hindbrain as well. For example, itis well established that the cerebellum is strongly aected in autismwith decreased numbers o the inhibitory Purkinje cells (Ritvo et al.,1986; Bailey et al., 1998; Bauman and Kemper, 2005), which wouldpredict hyper-reactivity. It is also known that there are increasedwhite matter volumes (Courchesne et al., 2001) which would

increase activation o brain regions rom other regions. Indeed,

others already postulated a disinhibition o the cerebellum thatwould lead to increased reactivity and could alter connectivity notonly within the cerebellum, but also across the cerebello-thalamocorticial circuits (Carper and Courchesne, 2000; Belmonte et al.,2004b; Boso et al., 2010).

In summary, autonomously acting hyper-unctional microcircuits in the neocortex may cause exaggerated perception to

ragments o a sensory world that must normally be holisticallyprocessed, and may cause hyper-ocusing on ragments o the sensory world with exaggerated and persistent attention. This hyper-attention could become dicult to shit to new stimuli due to thediculty or bottom-up and top-down mechanisms to coordinatethe overly autonomous low-level neocortical columns. The hyper-

plasticity component may also drive exaggerated memories to urther ampliy hyper-attention toward the same stimulus and driveover-generalization o attention to all related orms o the stimulus.The positive consequences are exceptional capabilities or elementary and specic tasks while the negative consequences are impairment o holistic processing, a rapid lock to a limited repertoire o

behavioral routines, which are then repeated obsessively.As a consequence the autistic person would remain with a rag

mented and amplied perception o bits and pieces o the world.The intense world that the autistic person aces could also easilybecome aversive i the amygdala and related emotional areas are signicantly aected with local hyper-unctionality. The lack o socialinteraction in autism may thereore not be because o decits in the

ability to process social and emotional cues, but because a sub-seto cues are overly intense, compulsively attended to, excessivelyprocessed and remembered with rightening clarity and intensity.Typical autistic symptoms, such as averted eye gaze, social withdrawal, and lack o communication, may be explained by an initialover-awareness o sensory and social ragments o the environ

ment, which may be so intense, that avoidance is the only reuge.This active avoidance strategy could be triggered at a very earlystage in a childs development and could progress rapidly with eachexperience maniesting as a regression, which is striking in somecases. With such early over-specialization, many other importantelementary certain skills may never be properly developed to enable

normal navigation in a socially rich world with an appropriate

understanding o social cues and communication.As already statedby other authors (or example Frith, 2003) autism aects development, and in turn development aects autism (p. 2). Compilinghigher order unctions such as abstract thought and language whenthe elementary alphabet o eatures is so overly attended to may

become dicult i not impossible in severe cases.

suPPort, contrAdIctIons, And unIfIcAtIon

The Intense World Theoryproposes that core elementary cognitiveconsequences in any child on the autistic spectrum are hyper-perception, hyper-attention, hyper-memory, and hyper-emotionality.



11/29


Hyper-capabilities are one positive aspect o such a brain, sensoryoverload, avoidance o stimulus-loaded situations and rapid lock-

down into behavioral routines are the downside o it. The nextour subsections will examine i there are indeed any indicationsor such postulates in the autism literature.

The remaining subsections will discuss support or the mainneurophysiological postulates o the Intense World Theory, namely

hyper-reactivity and hyper-plasticity.In each section we propose support or these postulates, point

out possible data contradictions and contradictory postulates oother neurological and/or psychological autism theories. Finally,the Intense World Theoryproposes to be a uniying theory o autismand as such will attempt to resolve apparent conficting data, interpretations, views and hypotheses, and reconcile debates.

hyPer-PercePtIon

Super-charged microcircuits in primary sensory areas could produce enhanced sensitivity to sensory stimulation and sensory overload. It is oten reported that autistic children are hyper-sensitiveto touch, noises and visual input and react with temper tantrums,

extreme anxiety, and even panic when exposed to novel or stimulus-overloaded situations. Sensory stimuli that are bearable andnormal to a typically developing child may be unbearable to anautistic child. A person who is suering rom sensory overload willnaturally avoid situations that are unpredictable and lled withaversive stimuli, such as supermarket visits, social encounters etc.The hyper-sensory component o autism is or example epitomizedin the personal account o the well-known autistic Temple Grandin

(Grandin, 1996). Autistics also pay extreme attention to detail andcan notice the smallest changes in their environment. These everyday experiences are supported by a number o rather recent quantitative scientic studies, which point to the superior perormanceo autistics in a variety o auditory and visual tasks.

In the auditory domain, autistics exhibit enhanced discrimination between auditory stimuli (Bonnel et al., 2003), more accuratelocal target detection o auditory stimuli (Mottron et al., 2000)and diminished global intererence on auditory processing (Foxtonet al., 2003). Moreover, autistic individuals also exhibit impaired

pre-pulse inhibition (McAlonan et al., 2002; Perry et al., 2007). Inthis paradigm, an auditory soter pre-pulse reduces the reactivity toa second louder pulse, a well-known phenomenon which is calledpre-pulse inhibition and represents a orm o sensorimotor gating. Impaired pre-pulse inhibition in autistics suggests impairedinhibitory regulation o more complex orms o sensory processing

and may be a contributing actor to sensory overload.In the visual domain, autistics exhibit enhanced visual discrimi

nation capabilities (Plaisted et al., 1998; ORiordan and Plaisted,2001; ORiordan et al., 2001), aster target detection in eature andconjunctive visual search (ORiordan et al., 2001; Jarrold et al.,2005), more accurate local target detection (Plaisted et al., 1999),diminished global intererence on visual processing (Mottron et al.,2003), and enhanced orientation discrimination o rst-order grat

ings (Bertone et al., 2005).In general, autistics perorm better than controls on tasks that

avorthe detail-oriented,piece-mealprocessingo stimuli,whichgave rise to cognitive theories such the Weak Central CoherenceTheory (Frith, 1989; Frith and Happ, 1994; Happ, 1999) and

Enhanced Perceptual Functioning Theory (Mottron and Burack,2001; Mottron et al., 2006). For example, autistic individuals perorm better on the Wechsler Block Design task, which is due to agreater ability to segment the whole design into its component parts(Shah and Frith, 1993), the Embedded Figures Test, which requiresto spot and reproduce a gure rom a complex background (Shah

and Frith, 1983; Jollie and Baron-Cohen, 1997) and graphical

reproductions o impossible gures (Mottron et al., 1999). On theother hand, autistics seem to perorm worse on tasks that requireholistic and contextual processing o stimuli (Happ and Frith,1997). These results were summarized in and emerged rom theWeak Central Coherence Theory, which claims that autistic perception is characterized by a ocus on details at the expense o eature

integration and holistic Gestalt-processing. Actually, it is postulatedthat a decit in holistic top-down processing o sensory inormation produces the advantage or local detail processing (Frith, 1989;Frith and Happ, 1994; Happ, 1999). Later it was suggested thatenhanced detailed-ocused perceptual processingper se, rather thana ailure in central processing, is the root cause (Plaisted et al.,2003) and the theory was revised accordingly more recently (Happ

and Frith, 2006). This is also the main hypothesis o the EnhancedPerceptual Functioning Theory, which claims that the main eatures

o autistic perception are locally oriented perception and enhancedlow-level discrimination (Mottron et al., 2006).

The Intense World Theory also predicts enhanced perceptiono sensory ragments, ocus on details and piece-meal perceptionand a decit in complex and more holistic processing and thereoreencompasses these dierent theories. The Intense World Theoryhowever explains integration decits dierently rom the WeakCentral Coherence Theory, namely as arising rom autonomoushyper-unctional neocortical columns that are more dicult tocontrol and orchestrate by both top-down and stimulus entrainment by bottom-up mechanisms, rather than a decit in top-down

pathways or mechanisms. The consequences are however the sameas those predicted by the Weak Central Coherence Theoryand theEnhanced Perceptual Functioning Theory.

hyPer-AttentIon

Attention is a conglomerate o dierent unctionalities, including sustained attention, orienting attention, response inhibition,and set shiting (Posner and Petersen, 1990). Sustained attentionis dened by the ability to maintain attention to repetitive stimuliover prolonged periods o times. Orienting reers to the capabilityto disengage attention, shit attention, and re-engage attention.Response inhibition reers to the capability to suppress irrelevant

or interering stimuli and nally, set shiting is thought to refect

cognitive fexibility and reers to the ability to shit to a dierentthought or action according to changes in a situation.

Autistics are well known to pay abnormal and obsessive attentionto detail, and to note and record their environment with exquisiteclarity. They can become hyper-ocused and locked-in on appar

ently arbitrary subjects o interest and sustain their attention onthese subjects or unusually long time periods. However, they arealso known not to pay attention to things on demand, or examplewhen pointed out by parents, being called by their name or whensomebody enters the room. It is in act notoriously dicult toengage their attention on demand. We suggest that this apparent



12/29


attention decit is the result o excessive on-going processing andexcessive attention to endogenous domains where attention is edback onto onesel. As a consequence o this internal hyper-ocus,it would be more dicult or another person to command theautistic persons attention as well as it would be more dicult or

the autistic person himsel to command his own attention voluntarily. This orm o exaggerated sel-engrossment with internally

on-going processes could perhaps also explain the apparent decitin theory o mind so oten reported in autism.

The scientic literature reports normal sustained attention inautism (Garretson et al., 1990; Buchsbaum et al., 1992; Siegel et al.,1995; Pascualvaca et al., 1998; Noterdaeme et al., 2001; Voelbel et al.,2006; Johnson et al., 2007), decits in orienting attention (Caseyet al., 1993; Wainwright-Sharp and Bryson, 1993; Courchesne

et al., 1994; Townsend et al., 1996; Wainwright and Bryson, 1996;Rinehart et al., 2001; Belmonte and Yurgelun-Todd, 2003; Renneret al., 2006), decits in disengaging attention (Landry and Bryson,2004),impaired set-shiting response, such as on the Wisconsin CardSorting test (Rumsey and Hamburger, 1988; Szatmari et al., 1989;Prior and Homann, 1990; Ozono et al., 1991, 1994; Ozono,

1995; Shu et al., 2001) and other set-shiting tasks (Geurts et al.,2004; Ozono et al., 2004; Kenworthy et al., 2005; Verte et al., 2005).Conficting results are reported concerning impaired response inhibition, which may be normal (Ozono and Strayer, 1997; Goldberget al., 2005) or impaired (Johnson et al., 2007; Luna et al., 2007). Theimpairments are usually attributed to an executive unction decitin autism mediated by an under-perorming prerontal cortex (or

review see Sanders et al., 2008).The Intense World Theorysuggests that hyper-unctional micro

circuits become autonomous processing modules that are dicult to

control voluntarily. Once activated these columns may reverberateand not require continual external stimulus entrainment and caneasily escape top-down control rom areas such as the prerontal

cortex. It is understandable that this has been interpreted as a decit in the prerontal unctioning, but in act the prerontal cortexmay even be over-perorming in its attempts to catch up with the

runaway columns. In act, we ound the same hyper-unctionalitycaused by hyper-connectivity and hyper-plasticity in the prerontalcortex as in the somatosensory cortex in the animal model o autism(Rinaldi et al., 2008a). The prerontal cortex may thereore actuallybe over-perorming, but relative to other activity in the neocortexonly appear as i it is under-perorming. Indeed, autistic subjectsexhibit hyper-activation in this brain area as revealed in recent MRIstudies (Gomot et al., 2008; Knaus et al., 2008; Dichter et al., 2009;Belmonte et al., 2010) and structural MRI studies most commonlyreport an overgrowth o ipsilateral cortico-cortical connections and

this overgrowth o white matter is most pronounced in the prerontal cortex (e.g., Herbert et al., 2003., 2004; Hardan et al., 2006; Craiget al., 2007). This is more consistent with a hyper-unctional prerontal cortex than weak long-range connections as usually interpretedrom MRI studies (Just et al., 2004, 2007; Koshino et al., 2005, 2008;Cherkassky et al., 2006; Kana et al., 2006, 2007).

hyPer-memory And Intellect

Enhanced learning and memory as proposed by the IntenseWorld Theory, could explain the astonishing savant talents, suchas exceptional memory or music, extraordinary calendar and

numerical calculations, the ability to draw complex scenes inexquisite detail rom memory, or extraordinary actual memory(Pring, 2005; Treert, 1999). However, the general tendency inautism research is to consider these exceptional cases almost asaberrations o autism only present in a small subsection othe autistic spectrum. Most (7075% o the cases) autistics are

usually classied as mentally retarded with diminished cogni

tive capabilities (American Psychiatric Association, 1980, 1987,1994, 2000; Lord and Spence, 2006). Within this ramework themost puzzling question is: can hyper-memory then be a corecognitive abnormality in all o autism? This chapter briefydiscusses the state o memory research in autism and makes a

statement regarding the general vista o low cognitive capabilities and mental retardation in autism. We argue that enhancedmemory capabilities in autism may lie at the heart o many coresymptoms o autism.

Initial scientic studies on autistic memory postulated a memory dysunction rather than hyper-unction behind the decits

in social interactions and language decits. In particular, the prevailing opinion was to compare autism to amnesia (Boucher and

Warrington, 1976). Later studies could not substantiate this claimand the current research vista states that basic perceptual basedmemory unctionality in the visual, auditory, spatial, and evenverbal domains is normal or even enhanced, but deteriorates with

increasing complexity and contextual enrichment (Minshew et al.,1992, 1997; Rumsey and Hamburger, 1988; Minshew and Goldstein,2001; Toichi and Kamio, 2002; Williams et al., 2005, 2006). Theconclusions is that an underlying core decit in Executive Function(Ozono et al., 1991; Bennetto et al., 1996) or a decit in ComplexInormation Processing(Minshew and Goldstein, 1998) contribute

to this type o memory dysunctions.The Complex Inormation Processing Disorder Theoryo autism

(Minshew and Goldstein, 1998) proposes an increasing impair

ment in integrating progressively more complex inormation dueto a ailure o neuronal integration mechanisms across dierentbrain regions, the so-called unctional UnderConnectivity Theory

(Just et al., 2004) and thus explains memory decits or complexand abstract material. While the Intense World Theoryalso predicts

decits in integration mechanisms, the latter theory suggests thatthe underlying neuronal mechanism is that o low-level hyper-unctional and autonomous columns that excessively process andstore simple eatures. Thus, the Intense World Theorypredicts thatsimple classical and operant conditioning mechanisms as well aslow-level and simple perceptual memory processing should be

enhanced in autism. Indeed, such enhanced memory capabilitieswere already observed in classical conditioning paradigms (Sears

et al., 1994) and some perceptual memory paradigms (Toichi andKamio, 2002; Caron et al., 2004).

However, the general tone o the vast majority o studies is tond and dene, with increasingly minuscule detail decits andimpairments in autistic memory and hypothesize about theircontributions to the autistic syndrome. Normal and in particular enhanced memory capabilities (Sears et al., 1994; Toichi and

Kamio, 2002; Caron et al., 2004) are usually ignored in neuropsychological theories o autism, with the notable exceptions o theEnhanced Perceptual Functioning Theory (Mottron et al., 2006;Mottron et al., 2010) and Extreme Male or Systemizing Theory



13/29


(Baron-Cohen, 2002). As noted above the current research vistastrongly associates autism with the stigma o mental retardationand low intelligence. However, is there really a oundation orthis association? Taking into account the proound consequenceso this type o stigmatization, it is a rather astonishing act thattypical diagnostic procedures or autism (such as the Diagnostic

and Statistical Manual o Mental Disorders, the Autism Diagnostic

Observation Schedule or Autism Diagnostic Interview) do notinclude a proper cognitive evaluation, but ocus on the symptomatic triad o social impairments, communication decits, andrepetitive behaviors (American Psychiatric Association, 1980,1987, 1994, 2000). Along this line, a meta-study evaluating 215

research articles published between 1937 and 2003 with claimsabout mental retardation ound that 74% o these claims camerom non-empirical sources, o which 53% were never traced backto empirical data. Astonishingly, most o the empirical evidencewas published 2545 years ago and based on measures o development rather than tests o intelligence or cognitive capabilities(Goldberg Edelson, 2006). But even when autistic intelligence istested with approved intelligence tests it is necessary to interpret

the results with caution. The most commonly used test to measure intelligence is the Wechsler Intelligence test and it was widelyapplied in autism research, yielding a characteristic prole o weakexecutive unction, low working memory, and low abstracting skills(Happ, 1994). Autistic people would usually perorm well only onone subscale, the Block Design test, which requires assembling ageometrical gure rom memory. In general, however, the WechslerIntelligence test is heavily based on verbal skills, command o lan

guage, and uses questions that require social and practical understanding. Based on this test, intelligence was generally classiedas low in autism with exceptional islets o perormance (Happ,1994). However, this view is now being seriously challenged roma number o quarters. A recent study compared perormance

o autistic children and adults on the Wechsler and the RavenProgressive Matrices Intelligence Test. This test measures high-levelanalytical reasoning, such as inerring rules, managing hierarchicalgoals, and orming high-level abstractions in a presumably nonverbal way (Raven et al., 1998). In comparison to the Wechsler Test,autistic subjects had higher intelligence scores on the Raven Testand did not dier rom control subjects, suggesting grossly under

estimated fuid intelligence and cognitive capabilities in autism(Dawson et al., 2007). Further support or under-estimated intelligence and abstracting skills stems rom the Extreme Male Theory,

which suggest that the prominent cognitive style is autism is pronetoward analyzing the variables in a system, deriving the underlying

rules that govern it and to construct, predict, and control systems

(Baron-Cohen, 2002). This cognitive style is called systemizingand reads like a conglomerate o higher order cognitive abstracting capabilities that autistics were previously believed to be lesscapable o. Research by Baron-Cohen and colleagues suggests thatautistics exhibit enhanced systemizing capabilities and evidencecited or this are savant talents, attention to detail, preerence or

rule-based, structured and actual inormation, higher scores ontests o intuitive physics, preerence or toys such as cars, obsessionwith collecting items, obsession with closed controllable system,such as computers, and enhanced systemizing quotients (reviewedin Baron-Cohen, 2002).

Accumulation o this type o evidence poses a serious challenge to the current mostly decit-oriented research approachesin autism. The Intense World Theorysuggests a novel strategy orautism research o cognitive capabilities and proposes to studyidiosyncratic excessive memory ormation patterns in autism.Probably due to enhanced sensory processing and hyper-attention,

the Intense World Theoryexplicitly predicts excessive and idiosyn

cratic memory ormation in autism. While it needs to be clariedi this hyper-memorization is a consequence o hyper-perceptionand hyper-attention or a super-capacity on its own, learning andmemorization patterns in autism should be clearly dierent romnormal children and even vary substantially between autistic

children, depending on their early lie experiences and exposureto sensory material. While the case study literature and anecdotal accounts were always numerous and rich in documentingunconventional learning strategies and unexpected mnemoniccapabilities in autistic children and adults oten discoveredonly accidentally by parents or care-takers (Baron-Cohen, 2003;Dawson et al., 2008), a recent large-scale survey o parents o over144 autistic children revealed that 43% o these children exhibited

exceptional memory capabilities or individually selective material.In up to 10% this pattern was even striking (Liss et al., 2006). It is

easy to argue that hyper-plasticity at the columnar micro-circuitlevel may account or these exceptional memory achievements aswell as the savant skills.

However, the Intense World Theoryalso draws attention to whatmay be a serious oversight in autism research and diagnosis: that isthe downside o excessive memory processing. Excessive memoryin low-level sensory and elementary cognitive regions could lead

to an early over-specialization o eature processing and misseddevelopmental opportunities to acquire a ull spectrum o low-level processing strategies and to build higher order strategies.This might lead to a ragmented alphabet o eature processing

capabilities in the vocabulary o sensory processing and impedethe development o higher cognitive unctions such as abstract

thinking and language processing. The autistic person may alsobecome locked into powerul memories that are dicult to corrector extinguish and that dominate every-day lie. Quick and almostarbitrary association building based on enhanced perception osensory eatures paired with excessive internal emotions positiveor negative may rapidly lock the person down into behavioral

routines. A ailure to extinguish such associations may underliethe insistence on sameness and obsession with routines and maymake rehabilitation dicult.

Finally, we would like to draw attention to ear memory ormation in autism, a domain which has not yet received enough atten

tion. While autism is clinically associated with enhanced anxietyand phobias (Muris et al., 1998; Gillott et al., 2001; Evans et al.,2005), the current scientic literature on ear memory ormationin autism is sparse, consisting o only two recent publications onhigh-unctioning autistic adults. Both o them report normal earconditioning (Bernier et al., 2005; Gaigg and Bowler, 2007) and thusstand in striking contradiction to the clinically observed enhanced

ears and anxiety. Based on the observed hyper-reactivity and hyper-plasticity in the amygdala and concomitant enhanced ear memoryormation in the VPA model o autism (Markram et al., 2008), weargue that urther research on ear memory ormation in autism



14/29


(such as studying younger age groups, using varied ear conditioning paradigms including controls or context) will provide insightsinto the underlying nature o withdrawal, social avoidance, andawkwardness.

hyPer-emotIonAlIty

The orth axis in the Intense World Theoryis proposed to be hyper-

emotionality as an inevitable consequence o limbic hyper-reactivity and hyper-plasticity. The amygdala, a key part o the limbicsystem, plays a pivotal role in modulating and regulating emotionalresponses (Davis and Whalen, 2001; LeDoux, 2003; Zald, 2003;McGaugh, 2004; Adolphs, 2006) and a malunctioning in this particular brain region has been proposed to underlie the social decits

in autism (Baron-Cohen et al., 2000; Sweeten et al., 2002; Amaral

et al., 2003; Bachevalier and Loveland, 2006; Schultz, 2005). Forexample, the Theory o Mind suggests that autistics are severelyimpaired in reading other peoples minds and empathizing withother people (Baron-Cohen et al., 1985). This theory involves twoelements: (1) the ability to attribute mental states to onesel andothers, to be able to distinguish between onesel and others and

realize that others have independent minds and may pursue dierent goals rom onesel; (2) the ability to express an appropriateemotional reaction to the other persons mental state, thus to beable to empathize with the others mind. The proposed decits inreading other peoples eelings and thoughts and the lack in empathizing with other people have been commonly used to explain theimpairments in social interactions and communication as well as

inappropriate responses in social encounters even in high-unctioning orms o autism such as in an Asperger. It was suggested thatthese decits are mediated by a not sufciently activatedamygdala(Baron-Cohen et al., 1999; Critchley et al., 2000; Pierce et al., 2001).These and other data, such as post-mortem examinations o amygdaloid morphology (e.g., Kemper and Bauman, 1998; Schumann

and Amaral, 2006), amygdala lesion studies in non-human primates(Bachevalier, 1994; but see also more recently, Emery et al., 2001;Prather et al., 2001), as well as comparison between amygdalalesioned patients and autistics (Adolphs et al., 2001) have led to

the Amygdala Theoryo autism (Bachevalier, 1994; Baron-Cohenet al., 2000; Sweeten et al., 2002; Amaral et al., 2003; Bachevalier andLoveland, 2006). In its current version it implies that the amygdalais hypo-unctioning, thus the autistic person does not eelenoughor does not process socio-emotional cues suciently (reviewed inMarkram et al., 2007b).

On the other hand there is evidence that the amygdala maybe overly activated in autism. First, structurally the amygdala isenlarged in autism as early as 18 months o age and this enlarge

ment persists throughout childhood until about 12 years o age(Sparks et al., 2002; Schumann et al., 2004; Mosconi et al., 2009).In adolescence the enlargement disappears (Schumann et al., 2004)and by early adulthood the amygdala may even end up smaller thanin control subjects (Aylward et al., 1999; Rojas et al., 2004). Thesechanges may refect an over-activation o the amygdala in early

childhood. Second, unctional hyper-reactivity was demonstratedwhen autistic subjects are conronted with socially relevant stimuli,such as aces and eyes (Dalton et al., 2005; Kleinhans et al., 2009;Monk et al., 2010). For example, Kleinhans et al. (2009) showedthat compared to controls the amygdala o autistic subjects exhibits

attenuated habitation to acial stimuli and that increased amygdalaarousal in autistics was associated with increased social impairment.Monk et al. (2010) recently showed that right amygdala activation isenhanced in autistic subjects during ace processing when controlling or attention, that is when the autistic subjects pay attentionto the stimuli. Dalton et al. (2005) revealed that high-unctioning

autistics showed greater activation in the right amygdala when

viewing amiliar and unamiliar aces and greater activation in thelet amygdala and also in the let orbito-rontal cortex when viewing emotional aces. Both areas orm part o the emotion circuito the brain and increased reactivity to aces in these areas meansa heightened emotional response to these stimuli. Autistics in this

study also spent less time xating the eyes region (deviant eye gazeis a core eature in autism American Psychiatric Association,2000). Moreover, in autistics, but not in controls, the amount oeye gaze xation was strongly correlated with amygdala activationwhen viewing both, inexpressive or emotional aces (Dalton et al.,2005). This suggests that that eye gaze xation is associated withemotional and possibly negative arousal in autistics and this couldexplain why autistics have trouble looking other people in the

eye. Eye contact and watching the acial expressions are one o therst signs o cognitively healthy inants, are natural to people, andserve to build the basis or successul navigation through a socialenvironment. For an autistic person however, these stimuli may be

just too intense or even aversive to cope with and hence they areavoided. Obviously, continuous avoidance o a special class o cueswill consolidate eature preerence processing and prevent learning

in this domain, thus some later developed social awkwardness andinappropriateness described in autism may be due to this lack oacquired knowledge. However, contrary to the decit-oriented ordisconnected Amygdala Theoryand Theory o Mindo autism, wepropose that the amygdala may be overtly active in autism, andhence autistic individuals may in principle be very well able to

attend to social cues, eel emotions and even empathize with othersor read their minds, but they avoid doing so, because it is emotionally too overwhelming, anxiety-inducing, and stressul.

The Intense World Theory proposes that amygdaloid hyper-reactivity and hyper-plasticity may in particular provoke a disproportional level o negative emotions and aect in autism, suchas elevated stress responses and anxiety as well as enhanced ear

memory ormation. Enhanced phobias and anxiety levels wererst noted by Kanner himsel in his original case studies (Kanner,1943) and later conrmed by population studies on children withautism (Muris et al., 1998; Gillott et al., 2001; Evans et al., 2005) andtheir relatives (Micali et al., 2004). A peek into the autistic world

o increased anxiety, stress, and ear ormation is delivered in the

ascinating introspection o autistics Temple Grandin and SeanBarron, who vividly describe how anxiety and ear lead to socialwithdrawal and avoidance (Grandin and Barron, 2005).

In the research community, the idea o generally enhancedarousal levels in autism was brought orward in the mid sixties(Hutt and Hutt, 1965; Hutt et al., 1965) and since then enhancedautonomic activity,in terms o either enhanced reactivity to stimula

tion, diminished habituation to stimuli or enhanced baseline levels,was reported in autistics using skin conductance or cardio-vascularindicators (Palkovitz and Wieseneld, 1980; James and Barry, 1984;van Engeland, 1984; Barry and James, 1988; Hirstein et al., 2001;



15/29


Ming et al., 2005) and increased stress responses were observedmeasuring stress hormone levels in the blood stream (Lake et al.,1977; Tordjman et al., 1997; Corbett et al., 2006). Interestingly, wehave also observed increased levels o the stress hormone corticosterone in the VPA-exposed rat ospring in the blood stream(unpublished data). Thus, based on the human subject studies and

our VPA rat model data the Intense World Theorysuggests that the

autistics perceive their surroundings not only as overwhelminglyintense due to hyper-reactivity o primary sensory areas, but alsoas aversive and highly stressul due to an overly reactive amygdala,which also makes quick and powerul ear associations with usually neutral stimuli ear o a color or example. A natural coping

strategy to deal with this kind o emotional overfow could be socialavoidance and withdrawal. In urther support o this view, decreasedamygdala activation has been linked to genetic hyper-sociability(Meyer-Lindenberg et al., 2005), whereas increased activation isobserved in social avoidance and phobia (Stein et al., 2002).

hyPer-connectIVIty

Autism was previously proposed as a disorder o brain connec

tivity (Belmonte et al., 2004a,b; Casanova, 2004; Just et al., 2004;Courchesne and Pierce, 2005a,b; Courchesne et al., 2005; Mottron

et al., 2006; Geschwind and Levitt, 2007; Minshew and Williams,2007). In human subjects, cortical connectivity can be studied eitheron the anatomical or the unctional level. White matter volumeswithin or between brain regions indicate macro-anatomical connectivity. Functional connectivity between brain regions is usually assessed as the amount o temporal activity synchronization

between these brain regions during task perormance. By comparing synchronization states o autistic versus control subjectsabnormal unctional connectivity patterns are deduced.

Magnet resonance imaging studies show that white mattervolumes are increased in young (23 years old) autistic children

in the cerebrum (18%) and cerebellum (39%) (Courchesne et al.,2001). This increase is not uniorm, but is most pronounced inthe rontal, ollowed by the temporal and parietal lobes, whereasoccipital lobes remain normal (Carper et al., 2002). In the same

two studies, older autistic children did not dier rom healthychildren in terms o their white matter volumes, suggesting aninitial overgrowth o connections, ollowed by an abnormally slowor arrested growth in later childhood and adolescence (Courchesneet al., 2001; Carper et al., 2002). Herbert and colleagues examinedwhite matter in more detail in 6- to 11-year-old autistic children.They subdivided cerebral white matter into an outer radiate zone

composed o intra-hemispheric cortico-cortical connections andan inner zone composed o sagittal and bridging compartments

mainly containing the long-range ipsilateral association bers,projection bers to sub-cortical regions and inter-hemisphericcommissural bers. They ound that while the long-range bersin the inner zone were not aected in the autistic group, the moreshort-and middle-range bers in the outer zone were increased in

volume in the autistic children. Consequently, they concluded thatshort- and middle-range connections are overgrown in autisticchildren (Herbert et al., 2004). Post-mortem studies indicate aMinicolumnar Pathologyin autism. Casanova et al. (2002) oundthat minicolumns are in general narrower in the rontal and temporal lobes o autistic brains than in normal brains. Both the

core, holding the cell bodies, as well as the neurophil, holdingunmyelinated axons rom inhibitory interneurons as well as dendritic arborizations and synapses, were ound to be reduced insize. Two important, albeit speculative, conclusions were drawnrom these observations: (1) An increased number o processing units as well as loss o inhibition in the neurophil could lead

to hyper-excitability (Casanova et al., 2002). (2) Increased num

bers o minicolumns could contribute to increased short-rangeco

the intense world theory – a unifying theory of the neurobiology of autism

Documents