Four years ago Thorkil Sonne realized that his young autistic son
possessed an extraordinary memory and a remarkable eye for detail. Those traits are prevalent among people with autism, and Sonne saw an opportunity to help individuals with the disorder find productive employment. As the technical director of a Danish software venture, he knew those qualities were critical in software testers. So he went out on his own and launched Specialisterne, a Copenhagen-based software-testing firm that now has 51 employees, including 37 with autism, and revenues of $2 million.
You started your company to improve the lives of people with autism. Why not just create a nonprofit focused on research or job training?
I wanted to do more than just provide a sheltered workplace for people with a disability. My goal is to create opportunities for people with autism on an international scale. You might find money to support sheltered working environments in Scandinavia but not in Poland or Spain or Brazil. To extend its reach, our organization needs the kind of funding that only a profit-making venture can generate. It must succeed on market terms.
Is it hard to reconcile two missions—serving customers and aiding people with a disability?
We’re constantly asked whether we support customers or a cause. We want to do both, of course, but we’re always fighting against the suspicion that we’re just a charity. Our corporate social responsibility profile might open doors with CEOs, but executives in charge of software testing aren’t evaluated on CSR, only on getting the most for the company’s money. To wipe away their suspicions, we must exceed performance expectations every time.
All our business comes from the private sector. Because Denmark has no tradition of social enterprises, the government doesn’t earmark contracts for companies like ours or give them tax breaks. We have to compete head on.
How does managing autistic workers differ from managing other people?
Most of our consultants with autism have a mild form called Asperger’s and are high functioning. Still, because they’re often hypersensitive to noise, they can be uncomfortable in open-concept office spaces without doors or walls. They also have trouble working in teams and understanding social cues, such as gestures, facial expressions, and tone of voice. You have to be precise and direct with them, be very specific about your expectations, and avoid sarcasm and nonverbal communication. Though we expect employees to do their jobs well, we don’t ask them to excel socially or to interact all the time with others. We just find them the right role. That takes tremendous stress off them. I think normality is whatever the majority decides it will be, and in our company people with autism are the norm.
What about the relationships between customers and your autistic consultants?
About 70% of our work is done at customer sites. The customer appoints a contact—someone who’s good with special people, who will select the right tasks and a comfortable place for them. We also give our clients a short introduction to autism and to our firm’s unique culture. After working with our consultants, the customers start being more direct with their own colleagues and stating their expectations more clearly. That’s helped them improve in an area that many companies struggle with.
In what other ways might firms benefit by adopting your techniques for managing autistic employees?
Companies sometimes unknowingly employ autistic people because the condition often goes undiagnosed. But people with autism aren’t the only employees who don’t thrive in open offices or in the traditional management system, with its emphasis on teamwork and unclear instructions like “Figure out on your own how to deal with this.”
You have to get the most from employees, especially when labor is scarce. Our sector is crying out for manpower, but Specialisterne has many job seekers knocking on the door. The key is to find situations that fit employees’ personalities and ambitions, not force everybody into one mold. That just causes stress, and workplaces already produce too much of that.
Cells taken from people with a rare syndrome linked to autism could help explain the origins of the condition, scientists suggest.
The Stanford University team turned skin cells from people with "Timothy syndrome" into fully-fledged brain cells.
The abnormal activity found in these cells could be partially corrected using an experimental drug, Nature Medicine reports.
UK researchers warned the findings might not apply to everyone with autism.
Compared with the hundreds of thousands of people worldwide thought to show characteristics of autism, "Timothy syndrome" is vanishingly rare, affecting an estimated 20 people across the planet.
People who have the syndrome frequently display autistic behaviour, such as problems with social development and communication.
Because it is caused by a single gene defect rather than a combination of small genetic flaws, each making a tiny contribution, it presents a useful target for scientists looking to examine what goes wrong in the developing brain of a child with autism.
Ready for work
The US researchers used a technique developed recently to generate brain cells called neurons from only a sample of the patient's skin.
This allowed them to examine their development in the laboratory, and even use them to test out possible treatments.
They found obvious differences between neurons grown from Timothy syndrome patients, and those from healthy "control" subjects.
The healthy neurons developed into different subtypes, ready for work in different regions of the brain.
In contrast, the proportion of neurons developing into each subtype was different in the Timothy syndrome samples - more were equipped to work in the upper part of the cerebral cortex, and fewer in the lower part.
This meant there were fewer neurons equipped to work in a part of the brain called the corpus callosum, which has the role of helping the left and right "hemispheres" of the brain communicate.
These differences echoed those already observed in mice specially bred with the Timothy syndrome genetic fault.
In addition, the neurons were making too much of a particular body chemical linked to the manufacture of dopamine and norepinephrine, which play a significant role in sensory processing and social behaviour.
Dr Ricardo Dolmetsch, who led the study, said that the abnormalities found tallied with other evidence that autism was due in part to poor communication between different parts of the brain.
The team managed to reduce significantly the number of these malfunctioning neurons by adding a drug as they developed.
This, they said, meant it might be possible one day to treat this defect in a real patient, although the drug used was not currently suitable for children due to side-effects.
The National Autistic Society gave a cautious welcome to findings, but warned that they did not necessarily offer insights into every form of autism.
Researcher Georgina Gomez said: "Timothy syndrome is only one form of autism and so these findings only give a very limited picture of what might cause the condition.
"More work would need to be done to substantiate this particular piece of research."
Babies born underweight are known to be prone to a variety of cognitive problems.
Babies born weighing less than 4lb (1.8kg) could be more prone to developing autism than children born at normal weight, a study suggests.
Writing in Pediatrics journal, US researchers followed 862 New Jersey children born at a low birthweight from birth to the age of 21. Some 5% were diagnosed with autism, compared to 1% of the general population. But experts say more research is needed to confirm and understand the link. Links between low birthweight and a range of motor and cognitive problems have been well established by previous research. But the researchers say this is the first study to establish that these children may also have a greater risk of developing autism spectrum disorders. The babies in the study were born between September 1984 and July 1987 in three counties in New Jersey. They all weighed between 0.5kg and 2kg or a maximum of about 4.4lb. At the age of 16, 623 children were screened for risk of an autism spectrum disorder (ASD).
Most low birthweight children don't have autism, and most children with autism don't have low birthweight.”
End QuoteProf Dorothy BishopUniversity of Oxford
Of the 117 who were found to be positive in that screening, 70 were assessed again at age 21.
Eleven of that group were found to have an autism spectrum disorder. From these results, the researchers calculated an estimated prevalence rate of ASD of 31 out of 623 children, which is equal to 5%. Jennifer Pinto-Martin, professor at the University of Pennsylvania School of Nursing and director of the autism centre where this research was conducted, said: "Cognitive problems in these children may mask underlying autism. "If there is suspicion of autism or a positive screening test for ASD, parents should seek an evaluation for an ASD. Early intervention improves long-term outcome and can help these children both at school and at home." Dig deeper But Dorothy Bishop, professor of developmental neuropsychology at the University of Oxford, said it was important to put the findings in perspective. "The association looks real, but nevertheless, most low birthweight children don't have autism, and most children with autism don't have low birthweight." Georgina Gomez, action research leader for The National Autistic Society, said more research is needed to confirm the link between low birthweight and autism and better understand why babies born underweight may be more prone to developing autism. "Low birthweight has been linked to a range of motor and cognitive problems and often goes hand-in-hand with premature birth and birthing complications. "It is important to dig down further to try to understand the biological processes and events that could explain this proposed connection."
New support for autistic people falling victim to crime
The card provides details for a personal contact police and care staff can approach
A new card will be available to people with autism to help them if they fall victim to crime or are involved in an accident.
The Autism Alert card carries contact details for a person who can help police, firefighters and medical and council staff give the best support. The National Autistic Society (NAS) Scotland has urged teenagers and adults with autism to apply for it. Northern Constabulary is among public bodies welcoming the card. NAS Scotland director Dr Robert Moffat said it was vital that emergency and care services had the right information to support people with autism. He said: "Being a victim of a crime or accident can be a stressful experience for anyone. "But for someone with autism, it can be particularly disorientating and frightening." Dr Moffat added: "People with autism often have difficulty understanding facial expressions, can be very literal in their understanding of questions and easily misinterpret others' intentions. "In an environment of serious crime or medical emergency these types of misunderstandings can have serious consequences." Highland Council, Highlands and Islands Fire and Rescue Service and NHS Highland have also offered their support to the initiative.
19 and functional magnetic resonance imaging (fMRI)20
Received 21 April 2011; revised 2 June 2011; accepted 3 June 2011
1Department of Psychiatry, Autism Research Centre, University of Cambridge, Cambridge, UK; 2Department of Psychiatry, Herchel Smith Building for Brain and Mind
Sciences, University of Cambridge, Cambridge Biomedical Campus, Cambridge, UK and 3MRC Cognition and Brain Sciences Unit, Cambridge, UK
Correspondence: Dr MD Spencer, Department of Psychiatry, Autism Research Centre, University of Cambridge, Douglas House, 18b Trumpington Road, Cambridge
Table 1 Main activations to happy and fearful versus neutral faces
MNI coordinates P-value
(FWE-corrected)
Z-score Cluster
size
Region
x y z kE (voxels)
Happy versus neutral faces
Control group
28 10 54 0.002 5.07 129 Left superior frontal gyrus
46 20 16 0.003 4.90 78 Right temporal pole
42 14 16 0.004 4.86 200 Left temporal pole
36 62 24 0.006 4.76 77 Left temporoparietal junction
54 64 10 0.009 4.66 115 Left posterior STS
44 52 28 0.010 4.65 60 Right FFA
4 26 54 0.012 4.60 64 Left dorsomedial prefrontal cortex
66 28 2 0.015 4.54 32 Right middle STS
28 92 8 0.018 4.51 26 Right cuneus
62 52 4 0.026 4.41 26 Left middle STS
24 94 8 0.031 4.37 20 Left cuneus
Sibling group
Nil
Autism group
Nil
Fearful versus neutral faces
Control group
44 48 22 0.006 4.78 38 Right FFA
Sibling group
40 42 16 0.005 4.80 33 Right FFA
Autism group
Nil
Abbreviations: FFA, fusiform face area; FWE, family-wise error; MNI, Montreal Neuroimaging Institute; STS, superior temporal sulcus.
Activated brain regions, corresponding MNI coordinates, cluster sizes, Z-scores and P-values. All analyses are corrected for multiple comparisons, and P-values are
expressed following whole brain level FWE correction at the threshold of Po0.05.
A novel functional brain imaging endophenotype of autism
MD Spencer et al
3
Translational Psychiatry
We conducted analyses of variance within PASW Statistics
18 to measure the overall effect of group on the primary
contrast activation data (happy minus neutral and fearful
minus neutral) for each region of interest. Age and sex were
modeled as covariates in all analyses. Similarly, we used
analyses of variance to investigate autism versus control,
control versus sibling and autism versus sibling differences,
again taking age and sex as covariates. We investigated
linear trend effects across the three groups using polynomial
regression, and where a statistically significant linear effect
was found, we examined the quadratic effect to confirm that
this was nonsignificant. We plotted the mean activation
contrast estimate (expressed in arbitrary units±standard
error of the mean) for the three study groups.
To investigate whether the atypical response to happy
versus neutral faces was driven by an atypical response to
happy or neutral faces, or to both, we examined the response
to faces versus fixation crosses. First-level analysis was as
above, taking the primary contrasts as happy and neutral
faces versus fixation crosses. Second-level statistical analysis
proceeded as described above for the emotional versus
neutral contrasts.
Results
Neural response to facial expressions of emotion: happy
versus neutral faces. We examined the differential
response within the brain to happy compared with neutral
faces. In controls, happy faces elicited increased activation
compared with neutral faces (Figure 1 and Table 1) within a
range of areas strongly implicated in face processing,
empathy and mentalizing: the right (P¼0.003) and left
(P¼0.004) temporal poles, left temporoparietal junction
(P¼0.006), left posterior STS (P¼0.009), right FFA
(P¼0.010), dorsomedial prefrontal cortex (P¼0.012) and
right (P¼0.015) and left (P¼0.026) middle STS. Increased
activation was also detected in the left superior frontal gyrus
(P¼0.002) and the right (P¼0.018) and left (P¼0.031)
cuneus. All P-values are expressed following correction for
multiple comparisons on a whole-brain level FWE basis. In
contrast, no activation differences were detected within
sibling and autism groups at the threshold of Po0.05 FWE
corrected.
To investigate biomarkers of familial risk compared with
autism versus control differences, we examined betweengroup
differences in the fMRI response in autism, sibling and
control participants within the specific brain regions identified
above as being significantly activated in controls to happy
versus neutral faces (listed in Table 1). For all 11 brain
regions, activation was significantly reduced in autism
compared with controls, with siblings demonstrating an
intermediate degree of impairment.
Activation in siblings was significantly reduced compared
with controls for 7 of the 11 brain regions: the left superior
frontal gyrus (P¼0.001; F¼11.664), the right (P¼0.002;
F¼9.986) and left (P¼0.005; F¼8.551) temporal poles, the
right middle (P¼0.004; F¼9.068) and left posterior
(P¼0.016; F¼6.064) STS, the left dorsomedial prefrontal
cortex (P¼0.005; F¼8.570) and the right FFA (P¼0.044;
F¼4.184) (univariate analyses of variance, covarying for age
and sex; Figures 1 and 2). Furthermore, for all 11 regions,
activation in the autism group was significantly reduced
compared with controls, the effect of group was significant
across all the three groups, and polynomial regression linear
contrast effects across all the three groups were significant
Middle
STS
Temporal
pole
FFA
-28 -16 +2
4.6
4.8
5.2
5.0
Superior DMPFC
frontal
Cuneus
Left Right
Posterior
STS
TPJ
P = 0.001 P = 0.002 P = 0.004
+10
0.2
0.1
0.0
0.3
-0.1
autism sibling control
autism sibling control
autism sibling control
autism sibling control autism sibling control
autism sibling control autism sibling control
autism sibling control autism sibling control
autism sibling control autism sibling control
contrast estimate +/- SE
0.3
0.2
0.1
0.0
-0.1
0.4
-0.2
contrast estimate +/- SE
0.2
0.1
0.0
-0.1
0.3
-0.2
contrast estimate +/- SE
0.3
0.2
0.1
0.0
-0.1
contrast estimate +/- SE
0.3
0.2
0.1
0.0
contrast estimate +/- SE
0.1
0.2
0.0
contrast estimate +/- SE
0.2
0.1
0.3
0.0
contrast estimate +/- SE
0.3
0.2
0.1
0.0
-0.1
-0.2
contrast estimate +/- SE
0.2
0.1
0.0
-0.1
0.3
-0.2
contrast estimate +/- SE
0.3
0.2
0.1
0.0
-0.1
contrast estimate +/- SE
0.3
0.2
0.1
0.0
-0.1
-0.2
contrast estimate +/- SE
+24 +54
P < 0.001 P < 0.001 P < 0.001
Left superior
frontal gyrus
Right temporal pole Right middle STS
Left posterior
STS
Left DMPFC Left temporal pole
P = 0.005
P < 0.001
P = 0.016
P = 0.002
P = 0.005
P < 0.001
Right FFA Left middle STS
P = 0.044
P < 0.001
P = 0.003 P = 0.007
P < 0.001 P = 0.001
Left cuneus
Left TPJ Right cuneus
Figure 1 Neural response to happy versus neutral faces. Activation differences
(means±s.e.m.) between the functional magnetic resonance imaging response to
happy and neutral faces in adolescents with autism (n¼40), unaffected siblings
(n¼40) and controls (n¼40). Activation map indicates neural response to happy
versus neutral faces in controls, and shows activations to happy versus neutral
faces (Po0.05, FWE corrected) overlaid onto the canonical Montreal Neurological
Institute (MNI) 152 template brain image (axial section, z-coordinate indicated in
Montreal Neurological Institute space), with the colored bar indicating the T-value of
the plotted activation differences. DMPFC, dorsomedial prefrontal cortex; FFA,
fusiform face area; STS, superior temporal sulcus; TPJ, temporoparietal junction.
A novel functional brain imaging endophenotype of autism
MD Spencer et al
4
Translational Psychiatry
with no significant quadratic component. For all 11 regions,
activation in the autism group did not differ statistically
significantly from activation in siblings (Table 2).
Neural response to facial expression of emotion: fearful
versus neutral faces. In controls and in siblings, fearful
faces elicited increased activation compared with neutral
faces (Figure 3 and Table 1) within the right FFA (controls:
P¼0.006, siblings P¼0.005; FWE corrected). However, the
autism group did not display any significant activation
differences at the threshold Po0.05 FWE corrected.
As with the happy versus neutral analyses above, we
examined between-group differences in the fMRI response in
autism, sibling and control participants within the right FFA,
characterized above as the brain region significantly activated
in controls in fearful versus neutral faces. Activation in the
autism group was significantly reduced compared with
controls and a significant polynomial regression linear
contrast effect with no significant quadratic component was
detected across all the three groups (Figure 3 and Table 2).
However, activation in the sibling group did not differ
significantly from activation in controls or the autism group.
Neural response to faces versus fixation crosses. These
findings demonstrate a clear linear progression across
autism, sibling and control groups for atypical fMRI
activation to happy versus neutral faces. To address the
question as to whether the neural basis for this marker is an
atypical neural response to happy faces, neutral faces or to
both, we used the same brain regions defined by the
significant activations within controls to happy versus
neutral faces (comprising the 11 clusters listed in Table 1)
0.2
0.1
0.0
-0.1
0.3
-0.2
autism sibling
Middle STS Superior frontal
gyrus
Right Left
Temporal pole Temporal pole
control autism sibling control
autism sibling control autism sibling control
contrast estimate +/- SE
0.2
0.1
0.0
-0.1
0.3
-0.2
contrast estimate +/- SE
0.2
0.1
0.0
0.3
-0.1
contrast estimate +/- SE
0.2
0.1
0.0
0.3
-0.1
-0.2
contrast estimate +/- SE
P = 0.004 P = 0.001
P = 0.005
P < 0.001
P < 0.001 P < 0.001
P < 0.001
P = 0.002
Figure 2 Differences between ‘unaffected’ siblings and controls with no family history of autism in the neural response to happy versus neutral faces. Activation differences
(means±s.e.m.) between the functional magnetic resonance imaging response to happy and neutral faces in adolescents with autism (n¼40), unaffected siblings (n¼40)
and controls (n¼40). Activation map corrected for multiple comparisons at Po0.05 family-wise error corrected, and overlaid onto a three-dimensional-rendered template
brain within MRIcron. STS, superior temporal sulcus.
Table 2 Between-group differences in activations to emotional versus neutral faces
Region of significant activation
in controls
Between-group differences
P-value (F-statistic)
Effect of group
(across all three groups)
Polynomial regression
linear trend effect
Control versus
sibling
Control
versus
autism
Sibling versus
autism
P-value
(F statistic)
P-value
Happy versus neutral faces
Left superior frontal gyrus 0.001 (11.664) o0.001 (17.222) NS o0.001 (9.448) o0.001
Right temporal pole 0.002 (9.986) o0.001 (13.703) NS o0.001 (8.994) o0.001
