Cognitive tests: interpretation for neurotoxicity? (Workshop summary).

The appropriate use and interpretation of cognitive tests presents important challenges to the toxicologist and to the risk assessor. For example, intelligence cannot be measured directly; rather intelligence is quantified indirectly by scoring responses (i.e., behaviors) to specific situations (problems). This workshop, "Cognitive Tests: Interpretation for Neurotoxicity?" provided an overview on the types of cognitive tests available and described approaches by which the validity of such tests can be assessed. Unlike many tools available to the toxicologist, cognitive tests have a particular advantage. Being noninvasive and species-neutral, the same test can be performed in different mammalian species. This enhances one's ability to assess the validity of test results. Criteria for test validity include comparable responses across species as well as similar disruption by the same neurotoxicant across species. Test batteries, such as the Operant Test Battery, have indicated remarkable similarity between monkeys and children with respect to performance of certain tasks involving, for example, short-term memory. Still, there is a need for caution in interpretation of such tests. In particular, cognitive tests, especially when performed in humans, are subject to confounding by a range of factors, including age, gender, and, in particular, education. Moreover, the ability of such tests to reflect intelligence must be considered. Certain aspects of intelligence, such as the ability to plan or carry out specific tasks, are not well reflected by many of the standard tests of cognition. Nonetheless, although still under development, cognitive tests do hold promise for reliably predicting neurotoxicity in humans.

Worksite behavioral research. Results, sensitive methods, test batteries and the transition from laboratory data to human health.

Ample evidence demonstrates that many chemicals, at some concentration, affect the nervous system. Disastrous incidents resulting from food adulteration, medical experiments and pharmaceutical errors, and industrial and environmental exposures warn of the serious potential for neurotoxicants to affect human health. This article identifies the approximately 185 epidemiological behavioral neurotoxicology studies published through 1989 of extended workplace exposures to chemicals. Approximately 250 different tests have been administered to exposed workers in primarily cross sectional studies, and statistically significant decrements were reported in 43% of the approximately 1100 test/population administrations (virtually all 185 studies employed several of the 250 tests). In this research, 28 different chemicals as well as multiple chemical exposures have been studied. The most extensive findings are seen in research on carbon disulfide, lead, mercury, and multiple solvent exposures, although three or more independent studies also have been reported on workers exposed to styrene and organophosphates. Analyzing the consistent findings in these studies, they reveal a broad spectrum of cognitive, motor, and affective or personality changes. The most frequently reported functional deficits are in tests of intelligence, memory, spatial relations, coordination, and speed plus coordination. Individual behavioral measurement methods and significant test batteries that have been employed to assess changes in human nervous system functions to determine the degree of risk posed by new and established chemicals are described; two of those human behavioral test batteries are predicted to dominate research at the onset of the 1990's. There is a lack of parallelism between the human test methods found in these batteries and the US Environmental Protection Protection Agency (EPA) and Food and Drug Agency (FDA) guideline methods for pre-production screening of chemicals and food additives in animals, and it appears unlikely that the widely employed human and animal behavioral test batteries, as currently evaluated, could identify disease complexes newly suspected of having a chemical etiology.

Animal test systems to study behavioral dysfunctions of neurodegenerative disorders.

The Eighth International Neurotoxicology Conference, Role of Toxicants in Neurological Disorders (1990), evaluated the evidence that chemical exposures may play a role in the development of neurodegenerative disorders. This article describes the major neurodegenerative disorders (Amyotrophic Lateral Sclerosis, Huntington disease, Parkinson disease, and Alzheimer disease) addressed at the conference, followed by a description of test systems or models developed to study behavioral aspects of these disorders in animals. However, due to the complexity of the disorders and the species in which they are found, fully-developed models in animals of neurodegenerative disorders are lacking. This suggests the need for a clear strategy for selecting behavioral tests in animals to study aspects of any neurodegenerative disorders. Such a strategy is here exemplified for Alzheimer disease (AD) as a prototypical neurodegenerative disorder. Since an animal model cannot provide the full range of effects of human neurodegenerative diseases, particularly AD which produces incompletely characterized cognitive deficits, a rodent model must at this time be drawn from multiple sources, including: (1) Tests currently used to identify in rodents deficits associated with AD; (2) tests to identify Alzheimer-related signs in patients; and, (3) tests that relate to theoretical constructs of human and animal cognition. A battery that draws from those sources could include tests of: (a) Spatial learning and memory (Morris Water Maze and Radial Arm Maze), (b) delayed recall match-to-sample; (c) serial response learning; and, (d) visual discrimination (e.g., vertical vs. horizontal stimuli). This battery will identify behavioral changes characteristic of early-, middle- and late-stage AD, afford the potential to relate the findings to theoretical constructs of cognition, and evaluate learning capabilities not previously studied in rodent models of neurodegenerative disorders.

Development of a neurobehavioral battery for children exposed to neurotoxic chemicals.

In recent years there has been heightened concern over the potential impact of environmental exposures on neurological function in children. Children are thought to be especially vulnerable to neurotoxic effects due to a number offactors including play behavior, differences in metabolism, and the development state of the brain. Performance tests from the computerized Behavioral Assessment and Research System (BARS) and other non-computerized tests have been combined to develop a brief battery that assesses multiple neurobehavioral functions in preschool children aged 4-6. Tests were selected to assess a variety of cognitive functions including attention, memory, motor speed and coordination and other executive functions. The battery has also been translated into Spanish and developed for use with Latino populations. Four to six-year-old children are particularly challenging because of the shorter attention span and lower motivation to complete an extended test session. When testing this group it is important to maintain the motivation of the child throughout the entire session in order to obtain accurate performance measures. A series of sequential pilot studies were used to select and develop appropriate methods and parameters for the tests in the battery. Although English-speaking children were able to complete the initial battery with minimal difficulties, several difficulties were encountered when the tests were administered to a Latino population. Cultural differences made some material inappropriate for testing due to unfamiliarity with the material, and in some cases items in a test had more than one correct translation which made administration difficult.

Establishing stable test performance in tests from the Behavioral Assessment and Research System (BARS).

Research to identify adverse effects in humans chronically exposed to neurotoxic substances in the workplace or environment typically assesses people at one point in time in a cross-sectional study. The most widely used strategy employs performance measures taken from a single point in time and compares these with either performance of a control group or established normative data. However, multiple comparison points of the same people on the same test allow the dissection of acute--from chronic--exposure effects, among other important questions. When performance measures are used from multiple points in time, within-subject deviations are examined. For either research design, the goal is to minimize the effects of practice and to obtain stable performance on a test. Demographic variables such as age, education, and cultural background or ethnicity influence performance on neurobehavioral tests. These variables may also influence the development of stable performance. Different populations may have different learning curves so that stable performance on a test is achieved with different amounts of practice. This is especially important when making comparisons across groups that may not have equivalent backgrounds. The performance of three groups, English-speaking adults, Spanish-speaking adolescents, and Spanish-speaking migrant adolescents, was examined. Each group completed a battery of neurobehavioral tests from the Behavioral Assessment and Research System (BARS) during four sessions. Repeated measures ANOVAs were used to investigate performance across time. Tests measuring motor performance produced stable performance from the first session. More complex tasks that involved attention and memory showed a practice effect across sessions.

Smiling faces and other rewards: using the Behavioral Assessment and Research System (BARS) with unique populations.

Concern about the exposure of children and adolescents to occupational and environmental hazards has increased, and so has the need to develop testing methods that can adequately assess the effects of exposure in children. A computerized testing system, the Behavioral Assessment and Research System (BARS), was successfully modified to test both younger populations and populations which do not speak English, the original language of the battery. These adaptations were modifications of the existing features of the BARS system which was designed to assess the broadest possible audience: simple language instructions broken down into basic concepts (step-by-step training with competency testing at each instruction step); a token dispenser along with a "smiling face" stimulus that reinforced appropriate performance; and adjustable parameter settings (e.g., number of trials, difficulty). Data from four groups demonstrate the feasibility of using BARS with children as young as preschool age and for non-English speaking children.

Lessons learned--15 years of the WHO-NCTB: a review.

Based on expert recommendations, the World Health Organization (WHO) recommended the Neurobehavioral Core Test Battery (NCTB) in 1983 to detect neurotoxicity in world-wide populations. The NCTB consists of 7 neurobehavioral tests (Digit Symbol, Digit Span, Benton visual memory test/recognition form, Santa Ana dexterity test, Simple Reaction Time, Pursuit Aiming II, and Profile of Mood States). Research with the NCTB provides the context for the results of a mini-symposium held in 1999 to discuss the lessons learned about using the NCTB in working populations. Speakers from Africa, Asia, Europe, and North America reviewed data from 94 studies using the NCTB, including 63 from China, 13 from Korea, 11 from Poland, three from South Africa, three from the USA, and one from Ecuador. Although a consensus was not sought in the symposium, the key lessons learned that emerged from the various presentations, were: * The NCTB is effective in testing adults with 12 or more years of formal education, but can not reliably test persons with less than 9 years of education. * People from cultures very different from those in Europe and North America (eg, aboriginal and African cultures) may not be tested effectively by the NCTB, although others such as at least some Asian populations, can be. To address this problem, the construct validity of the NCTB should be established in a wide range of countries. * The POMS measures of emotion proved to be very sensitive to neurotoxic exposures in several countries, but the POMS was the NCTB test most affected by cultural differences. The Digit Symbol or the Milan modification of that test was also highly sensitive to neurotoxic exposures. * Examiner drift following training to administer the NCTB is a significant problem in technically trained Examiners. Pursuit Aiming II is very difficult to score reliably.

Neurobehavioral evaluation of soil and structural fumigators using methyl bromide and sulfuryl fluoride.

Neurobehavioral functions affected by methyl bromide exposure were evaluated in California structural and soil fumigators using methyl bromide and sulfuryl fluoride. Sampling data revealed that structural fumigators are exposed for up to 1.5 hrs/day to 0-2.2 ppm methyl bromide and/or 10-200 ppm sulfuryl fluoride, and soil fumigators can be exposed to 2.3 ppm methyl bromide over an 8-hr day. Subjects were grouped for statistical analysis on the basis of exposure history: Those exposed primarily (80% or more of the work period with exposure potential) to methyl bromide (N = 32), primarily to sulfuryl fluoride (24), or to a combination of methyl bromide and sulfuryl fluoride (40-60% of each) for a minimum of one year (18), and those not exposed to high concentrations of any chemicals (29 Referents). Fumigators using methyl bromide reported a significantly higher prevalence of 18 symptoms consistent with methyl bromide toxicity than did Referents. Methyl bromide fumigators did not perform as well as Referents on 23 of 27 behavioral tests (chosen to reflect methyl bromide effects), and were significantly lower on one test of finger sensitivity and one of cognitive performance. These consistent differences suggest that even the low levels of methyl bromide found in fumigation today may produce slight neurotoxic effects. found in fumigation today may produce slight neurotoxic effects. The greater number of symptoms and reduced performance on all cognitive tests in sulfuryl fluoride fumigators compared to the Reference Group plus the absence of published research on this compound suggest that the data base for sulfuryl fluoride is inadequate.

Symposium on computerized behavioral testing of humans in neurotoxicology research: overview of the proceedings.

The Symposium on Computerized Behavioral Testing of Humans in Neurotoxicology Research, held in Portland, OR (USA), on June 21-23, 1995, is the subject of the 23 articles in this issue of Neurotoxicology and Teratology. It is the first open forum to focus exclusively on computer-implemented behavioral test methods for neurotoxicology. Both the Symposium and this proceedings have been organized around the topics: new technologies and new batteries, confounders and methodologic factors, special populations (children), sensory and motor testing, clinical applications, and neurotoxicology research. The Symposium emphasized common approaches to computerized testing, highlighted fundamental differences in strategies for the selection of tests, and produced tangible evidence that the NES2 test battery has become a pervasively used instrument for presenting valid tests in human neurotoxicology research. However, the dominant impression drawn from the Symposium is that there is a methodologic vigor in the field of computerized behavioral testing. This is revealed by the appearance of new test batteries, criticism of existing procedures, challenges to the existing order of test selection, and the identification of analytic covariates for commonly used tests.

Toxicity evaluation of sub-chronic exposures to cyanogen in monkeys and rats.

A 6 mon (6 hr/day, 5 days/week) inhalation toxicity study was conducted with cyanogen gas using male rhesus monkeys (Macacca mulatta) and male albino rats (Charles River Strain) as experimental animals. Fifteen monkeys and 90 rats were divided into three groups of 5 monkeys and 30 rats. One group, the Controls, was not exposed to the test material; the other two groups were exposed to either 11 ppm or 25 ppm cyanogen. At the outset of exposures, there was a doubling of the rate of responding on a variable interval 2.9 min schedule of reinforcement in monkeys exposed to 25 ppm cyanogen, and increases were also seen in the monkeys receiving 11 ppm exposures; the increases were transitory as the rate returned to control levels before exposures were terminated. At the end of the 6 mon exposure, there were no effects in hematologic or clinical chemistry parameters attributable to the inhalation exposure to cyanogen. The electrocardiograms, and gross pathologic and histopathologic examinations of test animals were normal when compared with the Control animals. Total lung moisture content was significantly lower in monkeys exposed to either 11 ppm or 25 ppm cyanogen than in Control animals. Body weights were significantly lower in rats exposed to 25 ppm than in Controls. The results suggest that subchronic 25 ppm cyanogen exposures are marginally toxic, but the evidence on 11 ppm does not support a similar conclusion.

Computerized neurobehavioral testing: techniques for improving test instructions.

The development of new computer-administered neuropsychological tests has brought a renewed focus on the issue of written instructions. Designers and programmers often fail to take advantage of the display options and feedback potential available in modern computer systems to maximize the effectiveness of the instruction process. Guidelines from computer interface architecture and behavioral analysis are combined to present three principles for written instructions: precise language, attention focusing, and interactive instruction. A comparison of different instruction formats in the Symbol-Digit and Simple Reaction Time tests shows that the application of these principles can reduce training time without degrading performance.

Human behavioral assessment in neurotoxicology: producing appropriate test performance with written and shaping instructions.

Neurotoxic effects are of such breadth and complexity that functional biomarkers (behavioral tests) that integrate many areas of the nervous system predominate in human neurotoxicology research. The increasing distribution of chemical and other manufacturing throughout the world, particularly in developing nations, suggests the acute need to develop biomarkers for chemical exposures and effects that can be employed internationally. A language-free method for training performance on behavioral tests is described, which holds promise for international research that circumvents the vagaries of translation. Four behavioral tests were administered to 74-114 adult US subjects. Procedures, collectively termed shaping, produced effective performance on three tests [Symbol Digit, Vigilant Attention Test (VAT), Digit Span Forward and Backward], and produced appropriate but unacceptably slow performance in initial testing on the Simple Reaction Time test. Effective performance on the Symbol-Digit test also was produced by shaping instruction, without assistance from examiners, in small groups of residents of Taipei (Taiwan) and US children between the ages of 5 and 16.

Adoption of an adult environmental neurobehavioral test battery.

Nationally recognized experts participated in a 3-day workshop to discuss the complex issues associated with neurobehavioral testing in environmental health settings, and to propose basic and focused test batteries for use in evaluating populations living near hazardous chemical sites. The Adult Environmental Neurobehavioral Test Battery (AENTB), which evaluates major neurobehavioral domains and functions, was adopted by the Agency for Toxic Substances and Disease Registry (ATSDR) for use as a basic screening panel in field studies. Pilot testing of the AENTB demonstrated an examiner training requirement of 3-6 practice sessions, a mean total testing time of 58.0 min (SD = 9.6), and, for 9 of the component tests, a sample size requirement of fewer than 140 (alpha = 0.05, 95% power) to detect a 20% difference between study groups. ATSDR administered the AENTB to 467 persons, selected randomly from 1,382 participants in field study sites in three states. Total testing time varied noticeably by participant age and study site, suggesting an ongoing need for site-specific controls in each field study using the AENTB. Also planned is adoption of a pediatric test battery to evaluate the domains and functions most relevant at major stages of child development.

Computer-administration of questionnaires: a health screening system (HSS) developed for veterans.

The introduction of microcomputers in psychological research has spawned a burgeoning number of tests of psychological or behavioral function, but few computerized systems for administering questionnaires have been developed. A Health Screening System (HSS) is described that combines the benefits of the paper-and-pencil format (e.g., convenient navigation within test questions) and the added benefits of computer-implementation (e.g., efficiency, automated scoring). The HSS features; a) appealing test appearance (e.g., text in large-size fonts, color backgrounds); b) clear wording of tests and instructions (identical wording as original tests except when clarity is served by changes); c) limiting need for Examiner-Subject interaction (e.g., continuously available on-line training, navigation within test questions, answer review capability, durable 9-button response unit); d) options (e.g., question skipping, spoken instructions, test questions, and answers on command); e) modification capabilities (e.g., color, text, test layout editing, control of test order, automated breaks, addition of tests to system); and f) extras (e.g., kernel of main instruction on each test screen, digitized video, audio message from Examiner in training, copyright notification on each screen, raw and summary data outputs in spreadsheet formal). Ten HSS tests were administered to 22 US military veterans, who took slightly longer to complete them than did 10 veterans who were administered the same tests in their original paper-and-pencil format. User reaction to the computerized HSS was positive.

Neurobehavioral test methods for environmental health studies of adults.

The Agency for Toxic Substances and Disease Registry convened a workshop in Atlanta, GA, that evaluated approaches and methods to ascertain whether there are neurobehavioral sequelae to children and adults exposed to hazardous substances in the environment. This article, developed from that Workshop, recommends testing methods [to identify neurotoxic insult] for immediate use in environmental health field studies of adults. A list of broad functional domains or modalities affected by chemicals was identified from the occupational and related literature (learning and memory, coding, sustained attention, higher intellectual function, strength, coordination, speed, vision, somatosensory, and affect). A core set of tests was selected that should assess those functions with the greatest demonstrated sensitivity to established neurotoxic chemicals, and a secondary set was selected to assess a broader group of functions. The core tests should be used in all investigations where neurotoxic effects would be targeted for identification; secondary tests would be used where suggested by questionnaire or symptom data or by knowledge of the effects of chemicals at the hazardous waste site.

Neurobehavioral effects of methyl bromide inhalation exposures.

Methyl bromide, a neurotoxic agent with a permissible US exposure limit of 20 ppm, is used primarily as a fumigant by an estimated 75,000 workers in the United States. This project was developed to evaluate the neurobehavioral effects of chronic and subchronic exposure to methyl bromide. One group of rats and rabbits was exposed to 65 ppm of methyl bromide for a total exposure of four 25-h weeks, or 100 h, and rats were exposed to 55 ppm of methyl bromide for a total exposure of 36 30-h weeks, or 1,080 h. Comparable control groups were given similar treatment, but no exposure. Behavioral tests of open field activity and limb coordination were conducted weekly during both phases of the experiment with rats, and eyeblink reflexes were measured weekly in rabbits. Nerve conduction velocity measurements were taken weekly from both rats and rabbits during the 65-ppm four-week exposures and monthly from rats during the 55-ppm 40-week exposures. Exposure to 65 ppm for four weeks significantly reduced eyeblink responses and nerve conduction velocity in rabbits but had no effect on rats. Exposure to 55 ppm of methyl bromide for 36 weeks had no effect on nerve conduction velocity, open-field activity, or coordination in rats.

Subjective cognitive complaints, affective distress, and objective cognitive performance in Persian Gulf War veterans.

We examined subjective cognitive complaints, affective distress, and cognitive performance in Persian Gulf veterans who reported illness and cognitive complaints. We predicted a stronger relationship between subjective cognitive complaints and affective distress than between subjective cognitive complaints and objective cognitive performance. This prediction was confirmed in a sample of 100 veterans. The results suggest that cognitive impairment should not be diagnosed in this population without objective confirmation with cognitive testing.

Comparison of MMPI-2 profiles of Gulf War veterans with epileptic and nonepileptic seizure patients.

As part of a larger study of illnesses related to service in the Gulf War, MMPI-2 profiles of epileptic seizure (ES) patients; nonepileptic seizure (NES) patients; Gulf War veterans with unexplained cognitive, psychological, musculoskeletal, fatigue, or dermatologic symptoms; and asymptomatic Gulf War veterans (Controls) were analyzed. There were 70 people in each group. Seizure diagnosis was based upon intensive EEG monitoring. Gulf War cases were mildly abnormal on MMPI-2 Scales Hs and D and significantly higher than controls on 8 of 10 MMPI-2 clinical scales, but they were significantly lower than NES patients on several scales including Hs and Hy.

Test-retest reliability of psychological and neurobehavioral tests self-administered by computer.

A series of 12 psychological and 7 neurobehavioral performance tests were administered twice to a nonclinical normative sample with 1 week between administrations. The tests were presented in a self-administered computerized format. One week test-retest reliabilities were comparable to conventional administration formats. The results suggest that individual test reliability is not affected when tests are administered as part of an extensive multi-measure battery. Computer administered test reliability coefficients also were compared to a Mixed Format (computer-conventional) administration with mixed format reliabilities generally similar to the reliabilities of published conventional tests but also generally lower than same format testing. Compared to psychological test reliability, neurobehavioral test reliability appeared more vulnerable to decreases with mixed format testing. These conclusions should not be generalized to all computer implemented tests as the qualities of the test implementation will affect the outcome.

The long-term effects of a token economy on safety performance in open-pit mining.

A token economy that used trading stamps as tokens was instituted at two dangerous open-pit mines. Employees earned stamps for working without lost-time injuries, for being in work groups in which all other workers had no lost-time injuries, for not being involved in equipment-damaging accidents, for making adopted safety suggestions, and for unusual behavior which prevented an injury or accident. They lost stamp awards if they or other workers in their group were injured, caused equipment damage, or failed to report accidents or injuries. The stamps could be exchanged for a selection of thousands of items at redemption stores. Implementation of the token economy was followed by large reductions in the number of days lost from work because of injuries, the number of lost-time injuries, and the costs of accidents and injuries. The reductions in costs far exceeded the costs of operating the token economy. All improvements were maintained over several years.