Related Topic(s):

Assessment of ADHD: A Neuropsychological Perspective

A chapter from the Textbook of Neurofeedback, EEG Biofeedback and Brain Self Regulation
edited by Rob Kall and Joe Kamiya

Assessment of ADHD: A Neuropsychological Perspective
Dennis Kade
Cumberland Hospital for Children and Adolescents
New Kent, Virginia


Research and clinical trials of neurofeedback for Attention Deficit Disorder (ADD) and Attention Deficit Hyperactivity Disorder (ADHD) have been criticized for several flaws. These include lack of rigor and specificity in the diagnosis of the primary condition, failure to assess comorbidity (which is highly related to prognosis) and use of outcome measures that are not accepted as valid in the broader field of ADD/ADHD research. Expanded assessment of these disorders is necessary to follow Lubar's recommendations for selecting appropriate patients for treatment, document generalization and establish credibility in the treatment of the Attention Deficit Disorders. A model is presented that can be used in the clinic or lab to assess individuals pre and post–treatment. It covers behavioral symptoms and cognitive abilities with known relationships to brain function. Researchers may use the model to pursue the relationship between measures of brain electrical activity and multiple diagnostic categories or variation in neurocognitive functions. The potential for new developments in the field of neuroscience are suggested from the wedding of neuropsychology and the psychophysiology of brain activity.

Assessment of ADHD: A Neuropsychological Perspective
Research and clinical trials of EEG biofeedback for individuals with Attention Deficit Disorder (ADD) and Attention Deficit Hyperactivity Disorder (ADHD) have been criticized for several significant flaws (Barkley, 1992). These include lack of rigor and specificity in the diagnosis of the primary condition, failure to assess comorbidity (which is highly related to prognosis) and use of outcome measures of questionable relevance (Barkley, 1992; Kade, 1993). The previously published studies (Lubar, 1991) have assessed treatment outcome with measures not accepted as valid in the broader field of ADD/ADHD research. Use of EEG data as outcome is circular reasoning. Data documenting EEG changes are important verification of adequate implementation of the treatment, analogous to measuring therapeutic blood levels of a drug in a pharmacological treatment study. However, EEG is a correlate of attention as blood pressure is a correlate of pain so that changes in EEG do not prove that significant change has occurred in the underlying condition. By analogy, changes in the EEG of a person with epilepsy are only of theoretical interest if there are no changes documented in the person's seizures. Also, changes in EEG have been produced by psychotherapy alone and thus offers no proof that EEG biofeedback is a necessary component of neurotherapy. Research documenting the efficacy of neurofeedback with ADD and ADHD have only recently begun to show an appreciation of these issues and still uses a paradigm that is out of date when compared to that used by researchers and practitioners providing more traditional treatment. We must expand our assessment of these disorders beyond brain electrical activity in order to document generalization and establish credibility in the treatment of the Attention Deficit Disorders. A more comprehensive assessment is also necessary in order to follow Lubar's recommendations (see his chapter in this book) regarding selecting appropriate patients for treatment. This chapter presents a model that can be used in the clinic or lab to assess individuals pre and post–treatment. Dimensions span behavioral symptoms and cognitive abilities with known relationships to brain function. Thus, the database is significantly expanded beyond assessment of brain electrical activity. Researchers may also use the model to pursue the relationship between measures of brain electrical activity and multiple diagnostic categories or variation in neurocognitive functions. It is both practical for the user and credible to those outside the field of biofeedback.

Differentiating ADD and ADHD: Core Neurobehavioral Concepts
If one considers the core behavioral deficit of ADHD as involving the "guidance system" of an individual, then the difficulties in starting/stopping and continuing in a behavior can be seen as two sides to a single neurobiological control system. Neuropsychologists would use the term executive functions to describe these behaviors and expect functional difficulties to be associated with problems in the frontal lobes of the cerebral cortex. The central behavioral complaints form two clusters, 1) poor persistence, seen as poor sustained attention or vigilance, and 2) poor ability to inhibit an initial reaction/response and/or physical movement. The first cluster of symptoms include impulsivity or a poor ability to delay gratification and hyperactivity. Thus, the individual with ADHD may drift off task due to poor persistence (symptom cluster 1) or be off task due to a reaction to some provocative aspect of the person's environment. The latter may be mistakenly labeled distractibility. A person with ADHD may behaviorally react to one part of the environment before taking time to fully scan the entire field or have considered all aspects (symptom cluster 2) or may have started off right, but then strayed from a course of action even though it was successful (symptom cluster 1). A third symptom cluster predicted by the neuropsychological model of this deficit could be described as a poor ability to organize a plan of action. Although not traditionally included as a core symptom, recent research has shown one example of this in that the individual with ADHD typically does not use the study skills that they have (O'Neill & Douglas, 1991).
In contrast, the individual with ADD can be expected to think before acting. They may fail to attend to all of the available information before acting but it seems to be due to a narrowness of scanning rather than the ADHD's rush to action in lieu of scanning. Indeed, individuals with ADD are sometimes described as lethargic and similar adjectives appeared on the attention–related factors of the empirically derived Child Behavior Checklist and Behavior Problem Checklist. The neuropsychologist will recognize the inattention of ADD as similar to the phenomena seen in individuals with deficits in the right posterior parietal region. It is not surprising that recent research has suggested that ADHD has a neuroanatomical basis in the frontal cortex and the frontal–subcortical white matter tracks. The research on ADD, as currently conceptualized, has not been done. However, previous research that probably included some ADD subjects suggested difficulties in visually scanning an entire field of information in some individuals with neurodevelopmental difficulties (Sebastofino, 1989; M. Morris, personal communication, April, 1990).
The following provides a comparison of ADHD and ADD, but it should be remembered that approximately 30% of individuals have both disorders (Barkley, 1993, January). There is also usually a difference in the family history, with a positive history of ADHD in individuals who have ADHD in a family history of learning disabilities and anxiety disorders and those with ADD [Barkley, 1990 #631].
Table 1


Poor Sustained Attention or Excessive Daydreaming,
Vigilance "Spacey" Appearance, Excessive Confusion or Mental "Fogginess"
Cognitive Sluggishness (Processing Disorder?)
Inconsistent Memory Retrieval?
Impulsive/Poor Delay of Able to Wait or Delay Gratification
Diminished Rule–Governed Rarely Impulsive, Disruptive,
Behavior Aggressive or Oppositional/Defiant
Hyperactivity/Poor Regulated Hypoactive/Lethargic Activity
Increased Variability of Task Possibly More Anxious than Normal or Performance ADHD
Socially Rejected Socially Reticent, Underinvolved, or Neglected

Note: 30% have ADHD and ADD. Adapted from Barkley (1992).

When considering clinical evidence of the above characteristics in assessing an individual, it must be remembered that the severity of the signs and symptoms of ADHD vary as a function of the situation or task. In addition to the factors (see Table 2) that tend to predict low vs. high severity of deficit, one must also be a close observer of the nature of the task.
One–to–one situations vs. group settings
Father's vs. mother's presence
Novelty vs. familiarity with tasks/settings/people
Frequent vs. infrequent feedback for performance
Immediate vs. delayed consequences for performance
High salience vs. low salience of consequences
Early vs. late in day
Supervised vs. unsupervised work performance

Note. Adapted from Barkley (1992)
For example, many parents are puzzled by the observation that a child with an attention deficit can play a video game for a long period of time or perhaps watch TV, but is completely unable to stay on task in the classroom for even 10 minutes. The casual observer does not consider that during 60 minutes of a video game, the number of "tasks" that a person has been attending to equals 60 multiplied by the number of times that task changed each minute. The parent who looks over the shoulder of their child will realize that the screen could, for example, change 4 times per minute and thus the child would be attending to 240 different tasks in that hour, not just one. Each one required only an attention span of 15 seconds. It is paced like Sesame Street when their classroom is paced more like Mr. Rogers' Neighborhood. The changing screen holds the attention. (Note that designers of neurofeedback displays should consider these factors. Many displays seem to hold a patient's attention for about three feedback sessions.)

After determining which symptoms are present, it should be documented whether the symptoms are exhibited across a wide variety of situations and tasks. This is seen as one index of severity and thus would be helpful for comparing populations included in different research studies. For example, in some countries other than the United States the only patients diagnosed with ADHD are those who exhibit global symptoms across situations.

DSM III–R and DSM IV Criteria for ADHD Across the Life Span
The important thing to note about the difference between DSM III–R and the proposed DSM IV Diagnostic Criteria for ADHD is that impulsivity is seen as the defining characteristic. It may or may not be accompanied by hyperactivity, but the individual is still considered to have ADHD rather than ADD.
Table 3
• Often fidgets with hands or feet or squirms in seat (adolescents, may be limited to subjective feelings of restlessness)
• Has difficulty remaining seated when required to do so
• Is easily distracted by extraneous stimuli
• Has difficulty following through on instructions from others (not due to oppositional behavior or failure of comprehension), e.g., fails to finish chores
• Has difficulty sustaining attention in tasks or play activities
• Often shifts from one uncompleted activity to another
• Often engages in physically dangerous activities without considering possible consequences (not for the purpose of thrill–seeking), e.g., runs into street without looking
• Must have at least 8 symptoms for at least 6 months since before 7 y.o.
Note. Adapted from Barkley (1992)
(Table Continued)

• Often has difficulty awaiting turn in group situations
• Often interrupts or intrudes on others
• Often blurts out answers to questions
• Often has difficulty playing quietly
• Often leaves seat
• Often runs about or climbs excessively
• Often fidgets or squirms
• Often talks excessively
• Often acts as if "driven by a motor" and cannot remain still
Must have > 5 for > 6 mo. since < 7 y.o.
If > 7 y.o. must also have > 6 mo. from ADD
Note. Adapted from Barkley (1992)
In the proposed DSM IV Criteria for ADD, this disorder regains the status that it had in DSM III after being relegated to "undifferentiated ADD" in DSM III–R.
Table 4

"Undifferentiated ADD"

Does not meet > 8 ADHD criteria for > 6 mo. since < 7 y.o.

Draft DSM–IV

1. Often has difficulty following through on instructions
2. Often has difficulty sustaining attention
3. Often seems not to listen
4. Often loses things necessary for tasks
5. Often fails to give close attention to details
6. Often is disorganized
7. Often makes careless mistakes in schoolwork or work
8. Often forgets
9. Often daydreams when should be attending
10. Often unmotivated to complete schoolwork or tasks.

Must have > 6 for > 6 mo. and < 4 Hyperactive–Impulsive symptoms

Note. Adapted from Barkley (1992)

It will be noted that the Draft DSM–IV criteria for ADD are similar at a behavioral level to neurological patients who fail to attend to all stimuli, showing what is termed inattention or neglect. Also compare these behaviors with the patients that Joel and Judith Lubar discuss (see Lubar's chapter in this book) regarding the patient with elevated theta power in the right posterior parietal region (detected by electrodes at P4–T6) and the hypokinetic or pseudodepressed patient.

When making a diagnosis of ADHD, it must be remembered that the diagnostic criteria are not presently adjusted for age differences. This is important since these behaviors are all frequent and normal in a young child but exceedingly rare in a normal adult. Barkley (1993, January) suggests that a cutoff of around the 93rd percentile will be appropriately established for 6 to 12 year olds using the DSM III–R guidelines. Suggested cutoffs across the age span are given in Table 5.
Table 5

> 10 if 2–4 y.o.
> 8 if 6–12 y.o.
> 6 if 12–17 y.o.
> 5 of > 18 y.o. (maybe 4 by late 20's)

Note. Adapted from Barkley (1992)

A cutoff on behavior rating scales used for diagnosis can be set at the 95th percentile (or the 93rd percentile at 1.5 standard deviations from the mean). These norms should be based on age and sex in order to establish the appropriate cutoff and mental age rather than chronological age should be used if the individual's standard score on a test of intelligence is below 85. It is also suggested that the criteria for symptom duration be increased from 6 months or more to a period of 12 months or more to avoid misdiagnosing some transient phenomena. This is particularly true for preschool aged children where there is a high rate of misdiagnosis. If the patient is an adult, then symptoms should be corroborated with another adult.

The patients should be subgrouped by comorbidity (coexistence) of Conduct Disorder or Anxiety (as well as by symptom pervasiveness as noted above) since these factors have been shown to be related to outcome for ADHD. Indeed, Conduct Disorder rather than ADHD alone may account for most of the negative adolescent and adult outcomes reported in early research that failed to document comorbidity of other diagnoses (Abikoff & Klein, 1992). This also assists in comparing results across research. For example, in Great Britain ADHD is not diagnosed if the criteria for a Conduct Disorder are met.

The Draft DSM–IV diagnostic criteria for Adult ADHD do not specify the need to rule out spectrum disorders, alcohol/substance abuse disorder, antisocial personality disorder, head injury and medical disorders. With changes in federal regulation that recognize attention deficits as potential disabilities, there is the increased possibility of secondary gains from the diagnosis. There is also the possibility that an individual might lie about signs and symptoms in order to obtain medication to abuse or sell. The classic closed head injury (see separate chapter) results in damage to frontal regions and associated EEG slowing. Thus, the individual may look as if they have one of the Attention Deficit Disorders and may show very similar behaviors.

Differential Diagnosis
If you simply give the parent or teacher a Conners Abbreviated Checklist and make the diagnosis based on a cutting score, you will be treating some misdiagnosed patients (Zelko, 1991). The scale is commonly used, in part because Abbot Laboratories widely distributed the scale and so many people are using the photocopy descendents as a free screening tool. If the patients diagnosed by elevated score on this checklist are then given Ritalin, a positive response was traditionally assumed to confirm the diagnosis. However, we know that both normal children and adults respond positively to mild stimulants by improved performance in many areas and thus a positive response is not diagnostic (Peloquin and Klorman, 1986; Rapoport, Buchsbaum, Weingartner, Zahn, Ludlow and Mikkelsen, 1980). The problem with using the Conners Abbreviated Checklist is that it does not allow a differential diagnosis; that is, it does not allow you to evaluate other conditions that could be producing the same behavioral symptoms. For example, an individual with a bipolar disorder will produce an equally elevated score on the checklist. Other issues include its sensitivity to aggression and lack of sensitivity to attention deficits (see (Atkins & Pelham, 1991) and Barkley, 1990, for reviews of this and other rating scales). Some of the competing explanations listed in Table 6 for the behaviors associated with ADHD are easy to rule out and some are difficult. Some are very rare conditions and some much more common. A good source of assistance in making many of these diagnoses (particularly of functional disorders) is Samuel, S. K. and Sikorsky, S. (1990).
Table 6
Organic Disorders

Lead intoxication
Sensory disorders, especially deafness
Frontal lobe abscess, neoplasm
Head Injury
Medication–induced attention deficit (e.g.,antihistamines, beta–agonists)
Substance abuse
Mental retardation
Seizure disorder
Learning disability
Pervasive developmental disorder
Sleep disorders

Functional Disorders
Conduct disorder
Oppositional disorder
Affective disorders with manic characteristics
Tourette's disorder/multiple tic disorder
Adjustment disorder with disturbance of behavior
Personality disorder
Anxiety disorder
Obsessive–compulsive disorder
Table 6 (Continued)


Age–appropriate overactivity

Inappropriate school placement: i.e., gifted child
in regular classroom, learning disabled or
developmentally delayed child in regular classroom

Chaotic home setting

Abuse or neglect or both

Note. Adapted from Garfinkel and Amrami (1992)

Some of the conditions may require screening by a neurologist, neuropsychologist, psychologist, or other appropriate specialist. Although at one point in the past lead intoxication was seen as a possible cause for all ADHD, more recent research has suggested that low levels are not significant [Barkley, 1990 #631]. Deafness may make a child seem inattentive. Acquired trauma required damage in the frontal lobes causes behavioral difficulties much like ADHD, presumably through damage to the same structures that are different in a child born with ADHD. In the same way that drugs are used to suppress the symptoms of ADHD, drugs can reduce symptoms of ADHD in a normal child. Because the behaviors that are associated with ADHD occur normally in individuals at a young age, the number of symptoms necessary for a diagnosis should be adjusted for age (see the previous section). If an individual has mental retardation, then his symptoms should be compared to his developmental age in order to make a valid diagnosis. Some seizure disorders produce a break in conscious mental activity as their only symptom. The difference is that a seizure will normally interrupt the ongoing flow of a person's behavior and the individual with ADHD is more likely to be described as "drifting" off task or shifting from one task to another rather than stopping in the midst of a task. It should also be noted that the Continuous Performance Test (CPT) (discussed below with other neuropsychological tests) was first used to evaluate inattention associated with absence seizures, once called Petit Mal epilepsy.

Individuals with a Learning Disability may be off task because they are frustrated with the academic task at hand. Also, individuals with a Learning Disability will often be distractible (Kupietz, 1990), pulled off–task by non–compelling stimuli within a preferred sensory modality. Careful reading of the initial description of ADD and ADHD in this chapter shows that children with ADHD are typically not distractible in this way. About 1 out of 20 children referred to an ADHD clinic may actually have a Pervasive Developmental Disorder (PDD) (Barkley, 1992, January). ADHD rating scales completed on individuals with PDD are elevated and thus a clinician using a one dimensional rating scale covering only ADHD symptoms will make a misdiagnosis. This is a significant error since individuals with PDD can decompensate on stimulants, even to the point that they require inpatient hospitalization. Conditions related to Pervasive Developmental Disorder include Autism, Asperger's Syndrome and Cohen's Multiplex Developmental Disorder. A screening checklist for the latter is available [Barkley, 1990 #631]. It is defined by a disorder in three or more of the following: thought (not delusions), affect, social, sensory and motor functioning. Typically there is a positive family history of psychosis or affective disorder. Diagnostic criteria for Asperger's Syndrome can be found in Firth (1990).

Oppositional Defiant Disorder and Conduct Disorder may be seen as a continuum of acting out. Individuals with these diagnoses may be out of their seat and performing poorly because of noncompliance. Indeed, the research that attempts to differentiate these disorders from ADHD at a symptom level is controversial. Attempts to differentiate these conditions using neuropsychological tests such as the CPT give inconsistent results ((Halperin, 1991)(Levy, Horn, & Dalglish, 1987)(Levy & Hobbes, 1989)(O'Brien, Halperin, Newcorn, Sharma, & et, 1992)(Power, 1992)(Trommer, Hoeppner, Lorber, & Armstrong, 1988)). Individuals with affective disorders with manic characteristics will meet the criteria for ADHD and score as high as a person with ADHD on Conner's Index. A careful examination will show, however, that they are cyclic in their symptomatology and the severity of their symptoms correlates with their mood. Although it is possible that individuals with Tourette's syndrome or Multiple Tic Disorder may be diagnosed with ADHD based on a superficial tallying of behavioral symptoms rather than an attempt at differential diagnosis (Biederman, Newcorn, & Sprich, 1991)(Comings & Comings, 1984)(Comings & Comings, 1990)(Golden, 1990)(Matthews, 1988). Over 70 percent of individuals with Tourette's are first diagnosed with ADHD. This is a significant problem since stimulants typically make tics worse. The anxiety disorders are also typically worse when placed on stimulant medication and so inaccurate differential diagnosis is important here, as well. One would not expect a person with an anxiety disorder to have problems with sustained attention. However, their immediate attention span may be worse than an individual with ADHD.

Other things to be considered include the high rate of misdiagnosis in preschoolers because of failure to adjust the criteria for number of symptoms and duration of symptoms beyond the standard DSM III–R guidelines. Thus, individuals with age appropriate overactivity may be misdiagnosed, only to "outgrow" their diagnosis by age 5 or 6. The possibility of inappropriate school placement should always be considered and though most evaluations might rule out a learning disability or developmental delay, my own clinical experience is that many clinicians evaluating for ADHD do not rule out giftedness as an alternative explanation for the behaviors. Abuse, neglect and a chaotic home setting can produce a child who is not prepared to positively adapt to the demands of a structured classroom and productive independent seat work.

Comorbidity: The things that can exist with ADHD
The possibility that an individual will have two different diagnoses, referred to as comorbidity or dual diagnosis has not been well studied in the attention deficits. We do know that there is at least a high rate of learning disability and Dr. Lubar's guidelines for neurofeedback suggest that this comorbidity may slow treatment response, but is not a contraindication for treatment. The same is probably true for conduct disorder and oppositional disorder. There is one published case study of neurofeedback for Tourette's syndrome, but no published neurotherapy with individuals having both Tourette's and an Attention Deficit. We know that there is also a significant rate of comorbidity with affective and anxiety disorders and these do seem to contraindicate neurofeedback.

It is just as important for the researcher to document comorbidity as for the clinician. Because more recent research (Hinshaw, 1987) has documented that the conduct disorder that is often comorbid with ADHD accounts for the most malignant outcomes that have been reported in literature on the prognosis of ADHD in adulthood, any treatment study of ADHD using any treatment technique needs to document comorbidity. Only in this way can the long term outcome be compared against a more accurate picture of the natural course expected for that individual. The discussion of prognosis naturally leads to considering the complications of ADHD.

Complications: Things that can develop from ADHD
The juvenile delinquency and chemical dependency that parents of ADHD children fear, is associated with conduct disorders rather than ADHD alone. Research which found these possible complications in adolescents was flawed by failure to determine comorbidity of the ADHD with conduct disorder in childhood. (Note that the research on family histories of individuals with ADHD as well as research on neurotransmitters which failed to isolate ADHD without conduct disorder are flawed; for example, more recent research suggests that the family history of pure ADHD is only positive for ADHD. See the review in Barkley, 1992.)

Academic underachievement is such a significant problem for the child with ADHD that the average individual must repeat two grades before High School graduation. The potential improvements in academic performance from neurofeedback are, in my opinion, their greatest promise. Traditional treatments, such as medication and behavior modification, have no documented improvements on academic achievement. Unlike the tendency for the primary symptoms to improve over time, academic achievement tends to decline (McGee & Share, 1988). Stimulant medication may increase productivity and task performance but there is no evidence of improved skill acquisition and achievement; the evidence is equivocal even when educational interventions are added to the treatment (Abikoff, 1991). The only other technique that has been shown to improve academic achievement is Satterfield's multimodal treatment (Satterfield, Cantwell, & Satterfield, 1979)(Satterfield, Satterfield, & Cantwell, 1980)(Satterfield, Satterfield, & Schell, 1987). This study was flawed by a very high dropout rate so that one cannot be sure whether the individuals did better because they stayed in treatment or whether they stayed in treatment because they were individuals who had a better prognosis to begin with and mistakenly attributed their disappearing symptoms to the treatment. Unfortunately, the researchers failed to include outcome data on the dropouts in their analysis of the treatment results. Ultimately, low self–esteem may be an even more significant complication than academic underachievement.

Once a person has finished with their formal education and finds a job that is compatible with their personality, interest, and previous training, their self–esteem is apt to be more important to their success. Here neurofeedback compares very favorably with medication since neurofeedback has been associated in reported case studies with improved self–esteem and there are concerns about the possibility that medication may contribute to low self–esteem by encouraging the child to attribute their successes to the medication rather than to their personal effort (Whalen & Henker, 1991).
Table 7


Academic underachievement
Low self–esteem
Oppositional behaviors/conduct problems
Negative interactions/relationships with peers, parents and teachers

Note. Adapted from Frick and Lahey (1991)

It is important to note that studies suggest that having 2 of the 13 DSM III–R Criteria for Conduct Disorder places a child at the 97th percentile and yet the criteria places the cutoff at 3 or more (Barkley, 1993, January). If this is followed strictly, then Conduct Disorders will be significantly underdiagnosed. It is also important to note that the behaviors listed are those typically associated with adolescents and so children may be undiagnosed. It is recommended that a multidimensional behavior rating scale be used to detect Oppositional Defiance and Conduct Disorders. The scale should include separate dimensions for ADHD. One excellent example is the Child Behavior Checklist. It comes in both parent (Achenbach and Edelbrock, 1983) and teacher forms (Achenbach and Edelbrock, 1986).

Conduct problems are a natural development for many individuals with ADHD and an important complication, even though they may not progress to the status of a comorbid Oppositional Defiant or Conduct Disorder. Edelbrock's developmental stage theory of childhood conduct problems is important to understand in this regard (See Table 7). Since one of the Stage 1 behaviors is impulsivity, all children with ADHD by definition are at least at some risk for developing conduct problems. The development of Conduct Disorders seems to a function of family genetics, degree of parental supervision and family dynamics (specifically a chain of coercive interactions between parent and child) and so family therapy should be recommended for an individual who is progressing into Edelbrock's Stage 2. There is consistent evidence of interpersonal problems in children with ADHD and a worse prognosis for those with peer problems, but assessment of social skills is somewhat complex (Atkins and Pelham, 1991). Ratings may be completed by teachers such as the Social Skills Rating System (Gresham and Elliot, 1990). Other approaches such as direct assessment via the individual's peers are preferable in many ways, but time consuming and potentially problematic. (For a review see ). This area is also important to assess for treatment effects. For example, Pelham and Hinshaw (1991) review review a study in which intensive behavioral and pharmacological treatment improved on–task behavior and academic performance but made no significant improvement in peer relationships.

Example Protocol Emphasizing Medical Aspects
Although the practice of obtaining a thorough physical evaluation of the child prior to taking a history and doing psychometric testing seems less and less a part of common practice, recent research suggesting that some individuals diagnosed with ADHD have a thyroid deficiency () may serve to renew its popularity. Garfinkel and Amrami (1992) present an assessment protocol for ADHD that includes a thorough evaluation of medical factors (see Table 8).

Table 8


1. Child Behavior Checklist
2. ADHD Rating Scale
3. Home Situations Questionnaire (<12 y.o.)
4. Conners Parent Questionnaire (?)
5. Issues Checklist (> 11 yrs.)
6. Interaction Behavior Questionnaire (> 11 yrs.)
Note: ADD may show as inattentive/passive on 1 or 4

1. Child Behavior Checklist Teacher's Rating Form
2. ADHD Rating Scale
3. Academic Performance Rating Scale
4. Conners Teacher Questionnaire (?)
5. School Situations Questionnaire (< 12y.o.)
Note: ADD may show as inattentive–passive on 1 or 5
Note: ADD may show as Inattentive/Passive on CBC or Conners

(Table Continued)

1. Child Behavior Checklist – Youth
or Behavior Assessment System for Children – Self Report
2. Issues Checklist
3. Interaction Behavior Questionnaire

Parent Self–Report:
1. Symptom Checklist – 90 Revised (SCL–90R)
2. Beck Depression Scale
3. Lock–Wallace Marital Adjustment Scale

Note: ADD may show as Inattentive/Passive on CBC or Conners. Adapted from Barkley (1992)
Note that this protocol should detect the organic disorders and the other issues listed previously when discussing differential diagnosis, but is less specific about ruling out functional disorders.

A different model with a more psychological focus is presented by Schaughency and Rothlind (Schaughency & Rothlind, 1991) and reproduced in Table 9. The advantage of this protocol is that it proceeds in a stepwise fashion so that the results from one part of the assessment may lead to a decision to discontinue assessment for ADHD and instead to pursue another avenue.

Table 9

Assessment Question Method Device/Tests

1. Does the child meet Structured diagnostic Diagnostic Interview for
the DSM–III–R diag– interview with multiple Children and Adolescents
nostic criteria for informants (DICA/DICAP)
Diagnostic Interview
Schedule for Children

Schedule for Affective
Disorders and Schizo–
phrenia for Children

Interview Schedule for
Children (ISC)

Child Assessment Schedule

(Table Continued)

Assessment Question Method Device/Tests

2. Does an alternative Structured diagnostic Same as 1
diagnosis or concep– interviews
tualization account
for his/her difficul– Other tests as appro– Tests of cognitive
ties (Differential priate (e.g., IQ, ability, Norm referenced
Diagnosis)? Achievement) achievement tests, cur–
riculum based assessment,
tests of adaptive

Information on Developmental history
developmental course
and onset

3. Does this child dis– Behavior rating scales Child Behavior Checklist
play these behaviors with multiple infor– (CBC)
to a developmentally mants
inappropriate extent Comprehensive Behavior
(for children with Rating Scale (CBRSC)
Mental Retardation,
compared to develop– Revised Behavior Problems
mental level)? Checklist (RBPC)

Conners Parent and Child
Rating Scales (PRS/TRS)
(Table Continued)

Assessment Question Method Device/Tests

4. Do these behaviors impair the Behavior rating scales Same as 3
child's functioning in the school, with multiple informants
in social rerelationships, and/or Sociometrics See Hops & Lewin (1984)
in the home?
Archival data –
referrals, grades,
classroom performance,

Note. Adapted from Schaughency and Rothlind (1991)

It is very important to note that the first step in this assessment protocol could be completed by a biofeedback technician with no training in clinical psychology or psychiatry. Structured interviews were developed to improve the reliability of diagnoses made using the DSM III–R and also for epidemiological studies in which a trained nonclinician would interview subjects. Shaunghency and Rothlind (1991) review the use of different structured interviews with ADHD and I would like to briefly summarize their review (see also Table 9). The DICA and DICAP include questions grouped by functional area rather than by diagnosis and so encourage a thorough review rather than just focusing on preconceived notions about specific diagnoses to be considered. They can be completed by a trained nonclinician and either hand or computer scored to produce a diagnosis. They have a known hit rate for ADHD: No false negatives but 22% false positives for data provided by parents and 50% false positives for data provided by children. These false positives rates can be improved by including data from the child's teacher obtained by rating scales (discussed below). The DISC and the parent version (DISCAP) can also be done by a trained nonclinician and are organized by functional area. On these scales, however, a clinician must review the results before scoring by computer.

Both the K–SADS and the ISC groups questions by diagnosis and requires that a clinician rather than a computer make the diagnosis. The K–SADS must be administered by a clinician. The ISC is similar to the K–SADS, but includes verbatim questions to be asked of the informant and thus can be used by a trained nonclinician.

The CAS must be done by a clinician but then is hand or computer scored. Questions are asked following rapport building and are grouped by functional area.

More recent developments allow the administration of the questionnaire by a computer on line and include procedures appropriate for diagnosing adults.

Rating Scales
Without a doubt, rating scales are the most widely used technique for both the initial evaluation of the Attention Deficit Disorders or the measurement of treatment effects. Their popularity can be attributed to their objectivity as well as their cost–effectiveness. As noted in the psychological assessment protocol above, they answer the question of whether the child displays these behaviors to a degree that is abnormal for their (developmental) age. Table _ lists four different rating scales that might be used to answer this question. A comparative review can be found in Schaughency and Rothlind (1991) and in Barkley (1991). The most important thing to consider in choosing a rating scale beyond the usual psychometric considerations of reliability and validity is whether it is multidimensional. Not only will this assist in differential diagnosis but allows the detection of a tendency on the raters' part to endorse any and all problems that might produce a false positive on a one dimensional rating scale of ADHD symptoms. At this time the best combination of rating scales to use in initial evaluations would seem to be those collected and developed by Barkley (1992) and listed in Table 8. It should be noted that the Child Behavior Checklist (CBCL) is an excellent multidimensional scale that comes in a parent, teacher and self rating forms. The teacher's ratings were traditionally considered most strongly in making a diagnostic decision because the teacher is apt to have a better sense of what is normal for a particular age because of the numbers of children of that age that the teacher has seen over his/her career. Using the same logic, a parent who is rating an only or an oldest child may not have a good base of experience on which to make ratings as to what is normal for that age. A child or adolescent with one of the Attention Deficit Disorders is unlikely to rate themselves as having any significant problems. Thus, the rating scales that they complete are more an aid in determining the presence of other problems. The Behavior Assessment System for Children (BASC) – Self Report scale is an alternative to the CBCL–Youth. Although the BASC is new and does not have the track record of the CBCL–Youth, it has the advantage of containing validity scales and covering many more dimensions that are clinically relevant, including self–esteem. The CBCL–Youth is intentionally a replication of the CBCL dimensions that are on the parent and teacher rating forms so that it can be determined if the adolescent sees him or herself in the same way as the parent and the teacher. In contrast, the BASC items and dimensions are based on an assessment of feelings and attitudes that only the child or adolescent may be aware of and able to rate. Barkley's ADHD rating scale can be completed by both parents and teachers and allows a normative interpretation of the number of ADHD symptoms that the individual is exhibiting. The Home Situations Questionnaire and the School Situations Questionnaire determine the pervasiveness of the difficulties (behavioral difficulties on the original form and attentional difficulties in the revised form). As noted above, pervasiveness is one index of severity. Also, in some countries such as Great Britain, pervasive symptomatology is required in order for the individual to be diagnosed with ADHD. The parent and teacher versions of the Conners Questionnaire may be preferred over the CBCL if the clinician knows at the outset that medication will be used as an intervention and the abbreviated form of the Questionnaire used to monitor treatment affects. The Issues Checklist and Interaction Behavior Questionnaire may be useful for the adolescent and parent to complete in order to gauge the presence of a significant parent–child conflict that could require separate treatment via family therapy. These measures could be repeated to gauge treatment benefits. As an alternative to the teacher's form of the abbreviated Connors Questionnaire, the Child Activity/Attention Profile can be used to assess treatment affects after initial evaluation with the CBCL Teachers Rating Form since it is a subset of the items of that multidimensional scale. If the teacher completes the Academic Performance Rating Scale the results can be used to objectively document the effects of the child's condition on academic performance. It can also be used to assess gains in academic performance due to treatment.

It is uncommon for clinicians assessing a child patient to systematically assess that child's parent(s). As suggested in the table, the SCL–90R will provide a general screening of the parent's psychological status in the week prior to the evaluation. The Lock–Wallace marital adjustment scale will give an overview of the marital functioning of the patient's parents. Any problems detected by these rating scales would be followed by additional evaluation. Parents should be informed of the reasons for obtaining this information and whether the results will be included in a report about their child. I have found parents to be quite understanding of the fact that their own personal or marital difficulties will affect their child to some extent and the rating scales will help the clinician to factor that in. It would also lead to additional treatment recommendations if problems were detected.

There is a long history of attempts at differential diagnosis and objective quantification of the Attention Deficit Disorders using psychological tests. The interested reader is referred to Richard Gardner's (1979) text entitled "The Objective Diagnosis of Minimal Brain Dysfunction." The difficulty with using such techniques for differential diagnosis is the high number of both false positives and false negatives. In the case of the former, virtually all of the traditional psychological tests that were used were found to also give abnormal results for children who were Oppositional Defiant or Conduct Disordered without ADHD. On the other hand, a significant number of individuals with ADD or ADHD may prove to be normal on the tests. It is very unfortunate that the label Freedom from Distractibility was given to the third factor of the WICS–R by Kaufmann (1979). Not only is it not diagnostic of ADD or ADHD (Barkley, 1991), but it contains three of the four subtests in the ACID (arithmetic, coding, information and digit span) profile that is seen in a subtype of learning disability (Rourke et al., 1989). This has lead many uninformed psychologists to mislabel students with this learning disability as having ADD or ADHD. Some feel that this third factor of the WISC–R or WICS–III cannot be accurately labeled or even interpreted as a unitary factor (see the reviews in the special edition of the Journal of Psychoeducational Assessment, 1991). Barkley (1992) does report that the Coding subtest of the WISC–R/WISC–III or Digit Symbol subtest of the Wechsler Adult Intelligence Scale–Revised (WAIS–R) may be low in ADD but not ADHD. Even this single subtest is difficult to interpret because it involves a complex combination of abilities including visual memory and learning as well as paper and pencil speed. The interested reader is referred to Spreen (1991) for guidelines on how to partial out the memory/learning factor. It must also be remembered that this subtest is very sensitive to many types of brain dysfunction (see Barth & Lee, 1992). Barkley (1992) also reports that the Hand Movement subtest from the Kaufman Assessment Battery for Children may be low in ADHD. Although the normative data only extends up through age 12.5, Spreen (1991) notes that these norms can be applied to adults and reports data for a small sample of adults. This test is also somewhat complex in that it involves immediate memory for a visual sequence but also requires the child to produce a sequence of motor movements accurately in order to respond to the test. Given the frequent occurrence of motor incoordination in ADHD and the fact that motor planning involves the central frontal region of the cortex, it would not be surprising if further research proved that the deficit in ADHD on this subtest was more in the motor execution than in the visual memory component.

Historically the Matching Familiar Figures Test has been used to objectively measure impulsivity. However, other scientists have interpreted poor scores on this test as evidence of anxiety rather than impulsivity. Reviewers have also noted the tendency for other diagnostic groups to score poorly on this test and for many individuals with ADHD to perform normally. Barkley (1992) suggests that the test may measure cognitive impulsivity (or what may be called the correctness of one's snap judgments) rather than behavioral impulsivity. Barkley maintains that ADHD is associated with the latter rather than the former. Although I agree with Barkley, I do feel that this test can give the clinician an opportunity to observe how the child responds to getting the feedback that his or her first answer was wrong. If the test scores on speed of response and errors are interpreted conjointly then extreme scores produce four possible categories; fast and accurate responders, slow and inaccurate responders, fast and inaccurate (considered cognitively impulsive), and slow but accurate. The first category is interesting because it could indicate that the child is gifted (and gifted but bored students are often mislabeled as ADD or ADHD). The second category might suggest a visual perceptual learning disability (since the test is a visual match-to-sample task). The last is interesting since it could be an indication of perfectionism or obsessive features. Since some of the characteristics of the current conceptualization of ADD suggest slowed processing, it is conceivable that this test might see revived use in quantifying slow response time in ADD. Individuals with ADD might fall into either of the two slow categories.

Intellectual and achievement testing may be necessary in order to rule out a Learning Disability or a general developmental delay such as Mental Retardation or Pervasive Developmental Disorder. The latter diagnoses of course require additional evaluation. Mental Retardation requires the assessment of adaptive behaviro; there are several excellent instruments to choose from (see Sattler, 1992, for a review of these and other psychometric instruments). Autism can be screened efficiently using the CARS and then subsequently evaluated using the Autism Screening Instrument for Educational Planning (ASIEP). A "multiplex developmental disorder" can be screened using a scale included in Barkley's (1991) packet of rating scales This disorder is perhaps best categorized as an Atypical Pervasive Developmental Disorder, but a later age of onset might suggest childhood Schizophrenia. Asperger's Syndrome shares some characteristics with autism and may be genetically linked with it, but is a much milder disorder and thus may go undiagnosed. A rating scale for Asperger's Syndrome based on research diagnostic criteria is available from the author.

All the techniques of personality assessment may be potentially useful in the differential diagnosis of emotional and behavioral conditions in both children and adults. As discussed above, these can either mimic ADHD symptoms, coexist with ADHD, or develop as complications from the condition. The selection of these tests is based on the signs and symptoms presented by the individual and is beyond the scope of this chapter (see texts by Knopf, 1990, and Sattler, 1992). It should also be noted that an adolescent or adult may exhibit symptoms very similar to ADHD secondary to the effects of substance abuse. A variety of screening techniques for substance abuse are available, but a laboratory drug screen should be considered.

Neuropsychological Tests
Neuropsychology involves the study of brain–behavior relationships and thus neuropsychological tests are those in which test performance has a known relationship to brain function. Since the Attention Deficit Disorders are neurodevelopmental conditions with presumed differences in central nervous system anatomy, individuals with these conditions should presumably perform differently from the average on neuropsychological tests. However, this hypothesis has not generally been borne out. The Wisconsin Card Sorting Test requires the examinee to sort cards into different conceptual categories while receiving feedback about the accuracy of their responses. At unannounced intervals, the contingencies of the task change. Thus, successful performance requires some degree of abstract concept formation, hypothesis testing, and efficient responses to feedback so that unsuccessful strategies are dropped quickly and successful strategies are maintained. Chelune, Ferguson, Koon and Dickey (1986) showed that 85.4 percent of the children with Attention Deficits were correctly classified on the basis of their high level of perseverative errors and failure to maintain a successful strategy. The differences between those with ADD and normals were greatest at younger ages. Subsequent studies had difficulties obtaining significant results, but this could be due to the ages in the sample. The neuropsychologist who is interested in using this test should consult the manual by Heaton, Chelune, Talley, Key and Curtiss (1993). It both reviews the most current research and provides guidelines for the complex interpretation necessary.

The Stroop Color–Word Test involves a series of three tasks in which the patient first reads a series of words that are the names of primary colors, then names the colors of ink that are used to print a series of x's, and finally names the color of the ink used to print a word while inhibiting reading the word itself (since the word is the name of a color and it does not match the color of the ink used to print it). This is a deceptively simple task. Psychiatric patients tend to perform below average across all three subtests. At least the more common type of dyslexia will produce a poor performance on the initial task relative to the other type tasks. Someone with word finding difficulties may have a problem with rapid naming of ink colors on the second task. A problem with the third task is interpreted as a deficit in inhibition. This deficit would be associated with cognitive functions normally attributed to the frontal lobes of the cortex. Thus, the history of poor performance on this test for individuals with ADHD is entirely consistent with what we know about the disorder (being associated with frontal lobe abnormalities).

There has been some report of abnormal learning curves on tasks in which a person with ADHD is required to learn something by rote memorization. Some drug treatment studies have used word or picture pairs, but normative data across the age span is lacking for such tests. The Selective Reminding Test has several things to recommend it. It is available in different forms that cover children through adults (normative data is available in Spreen, 1991). Alternate forms exist that allow retesting to measure treatment gains. Also, since the selective reminding format is most often used to detect memory deficits in mild head injury, it is presumably sensitive to mild difficulties compared to other learning tasks in which the entire data set is repeated each time for the patient regardless of their prior performance. Barkley (1991) has reported that on the Selective Reminding Test individuals with ADHD may show better encoding into memory than independent retrieval. Presumably this would be seen in a performance on cued or multiple choice recall that was normatively superior to the independent recall, a pattern I have noted in my clinical practice.

The Continuous Performance Test (CPT) is another deceptively simple measure that could have an entire chapter devoted to the intricacies of performance interpretation. Indeed, even the context in which the test is given can alter the results (Power, 1992). The popularity of its use with ADHD is due at least in part to Gordon's development of a practical commercial test apparatus with cutoff scores. The test requires the person to simply press something to respond whenever they see a given letter or number on the display in front of them. The examiner can then calculate the correct "hits" and the "misses" in which the target letter appeared on the screen and the person failed to respond. False alarms are tallied for each instance in which the person pressed for an incorrect stimulus. The stimulus can be simple, such as every letter A, or more complex. For example a memory component can be introduced by using a target of every A that immediately follows an X. Since the task moves quickly, a snap judgment is required of the person. Although the snap judgment is a rather simple one for an individual who is fluent in identifying letters or numbers, a person with a learning disability (LD) or who is just beginning their formal education may make errors in visual discrimination. Indeed, the test may not discriminate well between ADHD and LD (Kupietz, 1990). The different forms of the test available varies significantly in length. Thus, Barkley (1992) attributes the difference in his 35 percent hit rate (i.e., correct diagnosis) for the Gordon and the 60 to 70 percent hit rate for Greenberg's version of the CPT to the fact that Greenberg's CPT simply lasts longer. It would seem difficult to interpret some versions that take as few as five minutes as tests of "sustained attention." Conners (1991) has an interesting variation in which the person is required not to press whenever the target is present. This could theoretically provide a better measure of inhibition, but no normative data or hit rates are available and thus his version cannot be recommended for use in a clinical setting. Again, it must be remembered that a normal performance on the CPT cannot be taken as evidence against the diagnosis because of the rate of false negatives.

Recurring Evaluations: Documenting Treatment Gains
One might hope that the measures that are useful for differential diagnosis and documentation of the associated psychological and neuropsychological deficits might also be used to document treatment gains. However, a measure may be useful in documenting deficits because of its sensitivity to even very mild dys