http://www.niaaa.nih.gov/AboutNIAAA/Interagency/Reports/EarlyChildhood.htm#Summ

Early Childhood
Neurobehavioral Assessment
for the
Differential Diagnosis of Fetal Alcohol Syndrome
and Alcohol-Related Neurodevelopmental Disorder

Bethesda Marriott Hotel
Bethesda, Maryland

March 8–10, 2000

The Interagency Coordinating Committee on Fetal Alcohol Syndrome

An Overview of Current Models of Executive Function and Working Memory in Autism and Other Neurodevelopmental Disorders

Loisa Bennetto, Ph.D.

Assistant Professor, Clinical and Social Sciences in Psychology

University of Rochester

Dr. Loisa Bennetto discussed executive function (EF) in autism and other neurodevelopmental disorders. EF deficits might explain the diagnostic characteristics and some of the functional problems of children with neurodevelopmental disorders. EF deficits translate well into supports and compensatory strategies for children with developmental disabilities. However, executive dysfunction may not be an appropriate candidate for a core deficit, because the EF model has weaknesses. But reviewing the lessons from EF in other neurodevelopmental disorders may be helpful to the FAS community.

"EF" is an umbrella term that describes goal-directed and future-oriented neurocognitive processes and behaviors. EF involves the ability to inhibit behavior, cognitive flexibility, attention control, monitoring, planning and organizing, and working memory. Dr. Bennetto emphasized the importance of working memory—the ability to hold information in memory while processing other information—as a major EF.

The EF model has weaknesses that have implications for diagnosis and research in neurodevelopmental conditions. EF deficits are present in a diverse set of disorders with different etiologies, including autism, ADHD, fragile X syndrome (FXS), schizophrenia, ARND, and early treated phenylketonuria. The presence of EF deficits in multiple disorders with different causes has raised questions about the discriminate validity of executive dysfunction as a core deficit.

In addition, some neurocognitive assessment tasks that measure EF lack sensitivity, which results in nondetection in some patients and incomplete detection of some EF deficits in others. Impairment in EF may differ by age. For example, in some studies EF deficits were not observed in very young children with autism. This issue is complicated by the lack of full EF development in typically developing children, perhaps showing that finding EF deficits depends on cortical maturity in both patient and control groups. Alternatively, EF tasks might measure different cognitive functions at different ages.

Impairment in EF may also differ by IQ level. Thus, group differences may vary depending on the IQ of the subject and control groups. These findings may be due to the high correlation between EF and fluid IQ subtests. Impairments in EF are an important functional deficit, so treatment models should address them even if they are not an appropriate core deficit for diagnostic purposes.

Dr. Bennetto explained the reasons for EF’s limitations as a core deficit.There are concerns related to EF’s definition and the identification of EF’s component processes. The organization of EF components is still not clear, as investigators have not determined whether the components are independent or hierarchical. Additionally, the relationship between EF components and neural structures is not clear. EF is thought to be related to the frontal lobes, but EF deficits do not necessarily indicate frontal lobe damage because the frontal lobes are interconnected with many brain regions.

Moreover, EF presents measurement challenges to researchers. It is difficult to theoretically dissociate EF components, and few tasks assess only a single EF component. It is also difficult to distinguish nonexecutive components during the performance of EF assessment tasks.EF tasks typically involve several nonexecutive processes, and differences in nonexecutive processes can affect performance on EF tasks and could change the pattern of performance across different groups. For example, a nonexecutive function, such as spatial processing, can influence performance on spatial working memory tasks that assess EF. Similarly, social skills may affect performance on the Wisconsin Card Sorting Test (WCST) when it is administered in person versus by computer. Thus, processes that are not related to EF also affect many tasks used to assess EF.

Dr. Bennetto added that many EF assessments do not display a normal statistical distribution and lack reliability and statistical consistency. Low levels of reliability imply that the assessment task might not discriminate among groups with neurocognitive deficits. Some groups of children with neurodevelopmental disorders demonstrate a ceiling effect, as the test does not discriminate individuals performing at the higher end of the scale. In addition, many children reach adult levels of performance on EF tasks by early adolescence, which further restricts the range of scores. Because many EF tasks involve responding to and using information from novel situations, they are also sensitive to practice effects in longitudinal studies. Therefore, the usefulness of EF assessments for discriminating among different neurodevelopmental disorders is uncertain.

Dr. Bennetto next reviewed several of the putative components of EF, and described some of the traditional and experimental tasks that are used to assess them. Flexibility and set-shifting involve switching from one idea, response, or context to another. Impairment in this component is often evident in perseverative behavior and concrete, rigid approaches to problem solving. Planning involves identifying a goal, organizing the steps needed to reach that goal, and anticipating and generating strategies to avoid future problems. Organization can also be measured in the present, by assessing how a subject encodes or organizes complex stimuli.

Dr. Bennetto discussed how poor organization might lead to later difficulties in recall. Working memory is an ability that is distinct from short-term memory, and involves online storage and processing of information. Finally, tasks of inhibition measure the ability to inhibit the processing of irrelevant details or to stop oneself from carrying out a well-practiced response when it is no longer adaptive.

Research on EF in autism has found:

  • Impaired performance on tasks of cognitive flexibility and set shifting
  • Impaired ability to plan and organize
  • Impaired verbal working memory
  • Generally intact ability on simple inhibition tasks

This pattern of executive processes may help explain some of the social and other deficits in autism. In addition, looking for patterns both within executive processes and between executive and nonexecutive processes might address the issue of discriminate validity.

Investigations that compare EF in different neurobehavioral disorders may improve the discriminate validity of EF tasks and help differentiate among disorders.Dr. Bennetto described a study comparing three groups with EF deficits: Children with autism, women with FXS, and children with ADHD. Age and IQ were matched for each group, and each had an appropriate control group. Investigators examined the performance of these women and children on the WCST, focusing specifically on perseverative errors. All three groups with a neurobehavioral disorder showed impaired performance relative to their own control groups. The investigators also examined two tasks that specifically assess verbal working memory and ability to inhibit a response. Individuals with autism and women with FXS displayed a significantly poorer performance on the working memory task than their control groups. In contrast, children with ADHD displayed poorer performance on the inhibition task compared to their control group.

Dr. Bennetto indicated that the autism research could provide important lessons to FAS investigators. Researchers should focus on assessment tasks that have within-subject manipulations of executive and nonexecutive components or, in other words, tasks that permit the researcher to control for the nonexecutive components of EF tasks. To overcome the limitations of EF as a core deficit, researchers should also focus on the subcomponents of EF, and identify tasks that permit predictions from EF to other domains to determine if there are consistent patterns in different neurobehavioral disorders. Finally, it is important that the EF deficits characterizing a particular neurobehavioral disorder be consistent across age and IQ. It is difficult to conduct valid longitudinal studies without tasks that measure the same function in different age groups, but such studies could make invaluable contributions to our understanding of neurobehavioral disorders.

Discussion

Dr. Vivian Faden commented that the concept of IQ matching is intuitively appealing, but IQ is affected by some neurobehavioral disorders. She suggested that investigators match for IQ, age, and mental age. Dr. Bennetto agreed that IQ matching is an important issue. For autism researchers, matching for verbal or performance IQ, depending on the specific task, is a critical decision.