Estimation of premorbid intelligence: combining psychometric and demographic approaches improves predictive accuracy

A Normative Reference vs. Baseline Testing Compromise for ImPACT: The CARE Consortium Multiple Variable Prediction (CARE-MVP) Norms

BACKGROUND

Sports medicine clinicians routinely use computerized neurocognitive testing in sport-related concussion management programs. Debates continue regarding the appropriateness of normative reference comparisons versus obtaining individual baseline assessments, particularly for populations with greater likelihood of having below- or above-average cognitive abilities. Improving normative reference methods could offer alternatives to perceived logistical and financial burdens imposed by universal baseline testing.

OBJECTIVES

To develop and validate the Concussion Assessment, Research, and Education (CARE) Consortium Multiple Variable Prediction (MVP) norms for the Immediate Postconcussion Assessment and Cognitive Testing (ImPACT).

METHODS

We developed the CARE-MVP norms for ImPACT composite scores using regression-based equations. Predictor variables included sex, race (white/Caucasian, black/African American, Asian, or Multiple Races), medical history [attention-deficit/hyperactivity disorder (ADHD), learning disorder (LD), prior concussion(s), prior psychiatric diagnosis], and an estimate of premorbid intellect (Wechsler Test of Adult Reading). CARE-MVP norms were first validated in an independent sample of healthy collegiate athletes by comparing predicted and actual baseline test scores using independent-samples t-tests and Cohen's d effect sizes. We then evaluated base rates of low scores in athletes self-reporting ADHD/LD (vs. non-ADHD/LD) and black/African American race (vs. white/Caucasian) across multiple normative reference methods (Chi square, Cramer's V effect size). Lastly, we validated the CARE-MVP norms in a concussed sample (dependent samples t test, Cohen's d effect size).

RESULTS

A total of 5233 collegiate athletes (18.8 ± 1.2 years, 70.5% white/Caucasian, 39.1% female) contributed to the CARE-MVP norms (development N = 2616; internal validation N = 2617). Race and WTAR score were the strongest and most consistent ImPACT score predictors. There were negligible mean differences between observed and predicted (CARE-MVP) baseline scores (Cohen's d < 0.1) for all ImPACT composite scores except Reaction Time (predicted ~ 20 ms faster than observed, d = - 0.28). Low score base rates were similar for athletes across subpopulations when using CARE-MVP norms (ADHD/LD, V = 0.017-0.028; black/African American, V = 0.043-0.053); while, other normative reference methods resulted in disproportionately higher rates of low scores (ADHD/LD, V = 0.062-0.101; black/African American race, V = 0.163-0.221). Acute (24-48 h) postconcussion ImPACT scores were significantly worse than CARE-MVP norms but notably varied as a function of concussion symptom severity.

CONCLUSIONS

Results support CARE-MVP norm use in populations typically underrepresented or not adjusted for in traditional normative reference samples, such as those self-reporting ADHD/LD or black/African American race. CARE-MVP norms improve upon prior normative methods and may offer a practical, simple alternative for collegiate institutions concerned about logistical and financial burden associated with baseline testing. An automated scoring program is provided.

Similarities in cognitive abilities in older couples: a study of mutual influences

Objective: Similarities between spouses in cognitive functions have been mainly explained by the assortative mating phenomenon and the convergence for age and education. The mutual influence between spouses is another explanation particularly relevant in the elderly population. Today, it remains difficult to determine whether cognitive similarities exclusively result from the convergence effect or from the mutual influence. Using a novel methodology, the present study aimed to assess the impact of the marital relationship on cognitive similarities among elderly couples.Methods: 1723 couples from the Three-City Cohort Study were classified in two groups of couples with homogeneous and heterogeneous age and education. We also constituted two groups of pseudo-couples by a random association of individuals, with homogeneous and heterogeneous age and education. Dyadic analyses were conducted in the four groups, regarding the similarities in lexicosemantic abilities, executive functions, memory and global cognitive functioning.Results: Similarities were found on lexicosemantic abilities both in mate-assorted couples and in couples heterogeneous in age and education but no similarity was found in pseudo-couples.Discussion: Beyond the convergence effect, the fact that the spouses co-construct their lifestyles may contribute to cognitive similarities in the lexicosemantic domain.

Investigating the Test of Premorbid Functioning (TOPF) in predicting Wechsler Abbreviated Scale of Intelligence - Second edition (WASI-II) scores in an Australian sample

Accurate prediction of premorbid functioning is important in neuropsychological assessment. We aimed to investigate the predictive accuracy of the TOPF and examine this word list at an item level against WASI-II scores, using Australian pronunciations. The sample of 219 healthy Australians were aged 18-82 years. Multiple regression analyses were used to replicate the TOPF and simple demographic models based on the US TOPF standardization. Rasch analyses provided a comparison of Australian, US and UK word order from the proportion of words pronounced correctly. The variance explained in WASI-II index scores ranged from R²=.12 (PRI) to .33 (FSIQ-2), which was approximately half that reported in the US standardization study. The accuracy of predicted WASI-II scores was also slightly less in our sample. Thirty-two words were out of place by five places or more compared with the US word order and 30 compared with the UK. These results add to concerns about the application the TOPF with norms developed in the US and UK in the Australian context. Clinicians are advised not to apply the five error discontinue rule when using the TOPF in the local context. Development of a more accurate word reading task for use in Australia is warranted.

Comparing the North American Adult Reading Test (NAART) and the Test of Premorbid Functioning (TOPF) to estimate premorbid Wechsler Adult Intelligence Scale - 4th edition FSIQ in a clinical sample with epilepsy

Estimating premorbid general cognitive functioning is an essential component to the neuropsychological evaluation process. The North American Adult Reading Test (NAART) is a method to predict premorbid general cognitive functioning based on word reading skills developed using the Wechsler Adult Intelligence Scale-Revised (WAIS-R), which is currently in its fourth edition (WAIS-IV). The Test of Premorbid Function (TOPF) was developed using the WAIS-IV, based on the same method as the NAART, to estimate premorbid intellectual ability. There is a paucity of research comparing estimates of premorbid general intellectual ability between the NAART and TOPF. This study evaluated the clinical utility of premorbid estimates of FSIQ derived from the NAART and TOPF in a sample of 101 patients with temporal lobe epilepsy (TLE). Differences between NAART-derived premorbid FSIQs and TOPF simple demographic predicted FSIQs were significant (p < .001) with large effect sizes. NAART estimated premorbid FSIQ (M = 104.04, SD = 8.42) was significantly greater than TOPF premorbid estimates (M = 99.83, SD = 9.26). Results suggested NAART-derived estimates of premorbid FSIQ may be more accurate than TOPF-based estimates, which likely underestimated premorbid FSIQ in this sample of patients with epilepsy. Limitations and future directions are discussed.

Demographically Calibrated Norms for Two Premorbid Intelligence Measures: The Word Accentuation Test and Pseudo-Words Reading Subtest

The Word Accentuation Test (WAT, Spanish adaptation of the NART) and the Pseudo-Words (PW) Reading subtest from the Battery for Reading Processes Assessment-Revised (PROLEC-R) are measures to estimate premorbid IQ. This study aims to develop demographically calibrated norms for these premorbid measures in a representative sample of the adult Spanish population in terms of age, education, and sex. A sample of 700 healthy participants from 18 to 86 years old completed the WAT and the PW Reading subtest. The effect of age, years of formal education, and sex on WAT total score, PW total score, and time to complete the PW task (PW time) were analyzed. Percentiles and scalar scores were obtained for each raw score according to nine age ranges and individual education levels. The results indicated a significant effect of age and education on the premorbid performance assessed, with no significant effect of sex. Age and education explained from 1.9 to 33.2% of the variance in premorbid IQ variables. Older participants with fewer years of education obtained worse premorbid IQ estimates. This premorbid IQ estimation decline started in the 56-65 age range for WAT total score and PW time, whereas it started in the 71-75 age range for PW total score. This study reports the first demographic-calibrated norms for WAT and PW Reading subtest for Spanish-speaking population. Even though the influence of age and years of education on premorbid IQ measures was confirmed, the PW Reading subtest showed to be more resistant to decline in elderly population than the WAT.

Dementia Screening Accuracy is Robust to Premorbid IQ Variation: Evidence from the Addenbrooke's Cognitive Examination-III and the Test of Premorbid Function

BACKGROUND

Scores on cognitive screening tools for dementia are associated with premorbid IQ. It has been suggested that screening scores should be adjusted accordingly. However, no study has examined whether premorbid IQ variation affects screening accuracy.

OBJECTIVE

To investigate whether the screening accuracy of a widely used cognitive screening tool for dementia, the Addenbrooke's cognitive examination-III (ACE-III), is improved by adjusting for premorbid IQ.

METHODS

171 UK based adults (96 memory service attendees diagnosed with dementia and 75 healthy volunteers over the age of 65 without subjective memory impairments) completed the ACE-III and the Test of Premorbid Function (TOPF). The difference in screening performance between the ACE-III alone and the ACE-III adjusted for TOPF was assessed against a reference standard; the presence or absence of a diagnosis of dementia (Alzheimer's disease, vascular dementia, or others).

RESULTS

Logistic regression and receiver operating curve analyses indicated that the ACE-III has excellent screening accuracy (93% sensitivity, 94% specificity) in distinguishing those with and without a dementia diagnosis. Although ACE-III scores were associated with TOPF scores, TOPF scores may be affected by having dementia and screening accuracy was not improved by accounting for premorbid IQ, age, or years of education.

CONCLUSION

ACE-III screening accuracy is high and screening performance is robust to variation in premorbid IQ, age, and years of education. Adjustment of ACE-III cut-offs for premorbid IQ is not recommended in clinical practice. The analytic strategy used here may be useful to assess the impact of premorbid IQ on other screening tools.

Theory of mind and central coherence in adults with high-functioning autism or Asperger syndrome

The study investigated theory of mind and central coherence abilities in adults with high-functioning autism (HFA) or Asperger syndrome (AS) using naturalistic tasks. Twenty adults with HFA/AS correctly answered significantly fewer theory of mind questions than 20 controls on a forced-choice response task. On a narrative task, there were no differences in the proportion of mental state words between the two groups, although the participants with HFA/AS were less inclined to provide explanations for characters' mental states. No between-group differences existed on the central coherence questions of the forced-choice response task, and the participants with HFA/AS included an equivalent proportion of explanations for non-mental state phenomena in their narratives as did controls.These results support the theory of mind deficit account of autism spectrum disorders, and suggest that difficulties in mental state attribution cannot be exclusively attributed to weak central coherence.

The Abilities Associated with Verbal Fluency Performance in a Young, Healthy Population Are Multifactorial and Differ Across Fluency Variants

Numerous variants of verbal fluency tasks exist within clinical and research domains that purport to measure "executive function." However, to date, there has been a paucity of research examining what specific abilities are measured by these tasks. In this study, the relationships between a select group of cognitive constructs and phonemic, semantic, alternating, and excluded-letter verbal fluency tests were examined in 93 young healthy individuals (aged 18 to 35 years old). Forward-selection multiple regression analyses were performed for each fluency task. Phonemic fluency was associated with verbal intellectual function and processing speed; semantic fluency was associated with working memory and semantic word retrieval; excluded-letter fluency was associated with processing speed; and alternating fluency was associated with semantic word retrieval. These results highlight verbal intellectual function, processing speed, and semantic word-retrieval contributions to verbal fluency performances. The main conclusion from this study is that the abilities associated with verbal fluency performance in a young healthy population are multifactorial and differ across fluency variants. These findings progress our theoretical understanding of what is measured by different verbal fluency tasks and will assist interpretation of performance.

Associations between cortical thickness and verbal fluency in childhood, adolescence, and young adulthood

Neuroimaging studies of normative human brain development indicate that the brain matures at differing rates across time and brain regions, with some areas maturing into young adulthood. In particular, changes in cortical thickness may index maturational progressions from an overabundance of neuropil toward efficiently pruned neural networks. Developmental changes in structural MRI measures have rarely been examined in relation to discrete neuropsychological functions. In this study, healthy right-handed adolescents completed MRI scanning and the Controlled Oral Word Association Test (COWAT). Associations of task performance and cortical thickness were assessed with cortical-surface-based analyses. Significant correlations between increasing COWAT performances and decreasing cortical thickness were found in left hemisphere language regions, including perisylvian regions surrounding Wernicke's and Broca's areas. Task performance was also correlated with regions associated with effortful verbal processing, working memory, and performance monitoring. Structure-function associations were not significantly different between older and younger subjects. Decreases in cortical thicknesses in regions that comprise the language network likely reflect maturation toward adult-like cortical organization and processing efficiency. The changes in cortical thicknesses that support verbal fluency are apparent by middle childhood, but with regionally separate developmental trajectories for males and females, consistent with other studies of adolescent development.

Estimation of premorbid cognitive function based on word knowledge: The Swedish Lexical Decision Test (SLDT)

In clinical neuropsychology, the present status of a patient is evaluated in relation to the assumed premorbid status. However, in Sweden, existing methods to assess premorbid status are far from optimal. In the present study, the design and evaluation of a Swedish Lexical Decision Test (SLDT) for premorbid global cognitive function (i.e., premorbid intelligence) is described. The design was based on the empirical finding that, in general adult population, word knowledge is strongly associated with measures of global cognitive functioning. Linear stepwise regression analysis demonstrated that SLDT findings accounted for 48% of the variance of global cognitive function as assessed by the Full Scale Intelligence Quotient (FSIQ) from the Wechsler Adult Intelligence Scale Revised (WAIS-R). Demographic variables alone accounted for 31% and a combination of SLDT results and demographics accounted for 60%. Psychometric properties are presented using data from 109 healthy individuals stratified according to age, gender, and level of education. In addition, a case of Alzheimer's disease is presented to illustrate the relationship between SLDT performance and cognitive function. Finally, the theoretical foundation for the relationship between word knowledge and global cognitive function is discussed.

Comparative Utility of Barona Formulae, Wtar Demographic Algorithms, and WRAT-3 Reading for Estimating Premorbid Ability in a Diverse Research Sample

Various Barona formulae, a WTAR algorithm based on demographic data, and WRAT-3 oral reading methods of estimating premorbid ability were compared in a diverse research sample of 119 subjects. These methods were correlated with one another and with a modified version of the Raven Standard Progressive Matrices. Descriptive data are provided to illustrate advantages and disadvantages of various methods of estimating premorbid ability when no formal intellectual testing is available. While predicting premorbid ability for individual subjects involves varying degrees of error, we found that the revised Barona formula was superior to the original formula for subjects at the upper end of ability level. When researchers have screened out learning disability and have subject samples with few individuals likely to be of superior premorbid intelligence, oral reading scores are a reasonable measure of premorbid ability. Otherwise, researchers are advised to use both demographic and oral reading methods to estimate premorbid ability.

Clinical Validation of the General Ability Index—Estimate (GAI-E): Estimating Premorbid GAI

The clinical utility of the General Ability Index--Estimate (GAI-E; Lange, Schoenberg, Chelune, Scott, & Adams, 2005) for estimating premorbid GAI scores was investigated using the WAIS-III standardization clinical trials sample (The Psychological Corporation, 1997). The GAI-E algorithms combine Vocabulary, Information, Matrix Reasoning, and Picture Completion subtest raw scores with demographic variables to predict GAI. Ten GAI-E algorithms were developed combining demographic variables with single subtest scaled scores and with two subtests. Estimated GAI are presented for participants diagnosed with dementia (n = 50), traumatic brain injury (n = 20), Huntington's disease (n = 15), Korsakoff's disease (n = 12), chronic alcohol abuse (n = 32), temporal lobectomy (n = 17), and schizophrenia (n = 44). In addition, a small sample of participants without dementia and diagnosed with depression (n = 32) was used as a clinical comparison group. The GAI-E algorithms provided estimates of GAI that closely approximated scores expected for a healthy adult population. The greatest differences between estimated GAI and obtained GAI were observed for the single subtest GAI-E algorithms using the Vocabulary, Information, and Matrix Reasoning subtests. Based on these data, recommendations for the use of the GAI-E algorithms are presented.

Cognitive function and information processing in type 2 diabetes

AIMS

To determine whether uncomplicated Type 2 diabetes is associated with impairment of cognitive function and information processing ability.

METHODS

Thirty-eight participants with uncomplicated Type 2 diabetes and 38 non-diabetic controls were studied. The two groups were comparable for age and premorbid intellectual ability, and did not have other medical disorders likely to affect cognitive function. An extensive battery of tests was administered which assessed different levels and domains of cognitive functions including verbal and visual memory, executive function, general mental ability and efficiency of information processing.

RESULTS

No significant differences were found between the diabetic and control groups on any measure of cognitive function or information processing. The performance on these tests was not associated with recent glycaemic control (assessed by HbA1c). Duration of diabetes, however, correlated significantly with poorer performance on several measures of verbal memory.

CONCLUSIONS

The results of the present study suggest that some aspect of Type 2 diabetes (as indexed by the estimated duration of the disorder) does relate significantly to cognitive function within the group with diabetes. However, other diabetes-related factors, such as macrovascular disease, hypertension and depression, may contribute more to previously observed cognitive decrements in Type 2 diabetes.

Postdicting verbal IQ of elderly individuals

We undertook a longitudinal crossvalidation of the Ryan and Paolo (1992) equation's ability to postdict Wechsler Adult Intelligence Scale-Revised (WAIS-R) Verbal IQ (VIQ) from National Adult Reading Test (NART) performance measured 5 years after VIQ scores were obtained, for a sample of 49 elderly normal individuals (mean age 71 years). Five-year interval postdiction accuracy agreed very well with the results of the original, concurrent study. Clinical utility is still limited, however, as VIQ must decline by 16.3 points for 95%-detection sensitivity. A new regression equation that utilizes a combination of NART errors and WAIS-R Vocabulary age-scaled scores (measured 3 years earlier) provided slightly better expected clinical sensitivity and accounted for 49% of the variance in VIQ scores.

Cerebral perfusion correlates of depressed mood

BACKGROUND

The spontaneous diurnal variation of mood and other symptoms provides a substrate for the examination of the relationship between symptoms and regional brain activation in depression.

METHOD

Twenty unipolar depressed patients with diurnal variation of mood were examined at 8 a.m. and 8 p.m. with neuropsychological measures, clinical ratings and single photon emission tomography (SPET). Brain perfusion maps were spatially transformed into standard stereotactic space and compared pixel-by-pixel. A parametric (correlational) analysis was used to examine the relationship between symptom severity and brain perfusion, both between and within subjects.

RESULTS

Global depression severity and an independent 'vital' depression factor were associated in subjects with increased perfusion in cingulate and other paralimbic areas. In addition there was a probable association between an increase in an anxious-depression factor and reduced frontal neocortical perfusion.

CONCLUSIONS

Depressive symptom changes are associated with metabolic changes in the cingulate gyrus and associated paralimbic structures.

Psychological and demographic correlates of glycaemic control in adult patients with type 1 diabetes

The relationship between an objective measure of glycaemic control (glycated haemoglobin (HbA1)) and personality variables was examined in two separate groups of adult Type 1 (insulin-dependent) diabetic patients. Study 1 included 121 patients, all of whom also had subjective self-reporting of treatment compliance assessed, while the first 57 patients had individual differences in intelligence, major dimensions of personality and forgetfulness documented. Study 2 examined 303 patients, all of whom had their major dimensions of personality assessed using a shortened and updated version of the original personality questionnaire. Demographic indices (age, onset-age, duration of diabetes) were assessed in both groups. No significant correlation was found between HbA1 and self-report compliance suggesting that self-reporting may be invalid as a measure of glycaemic control. In study 1 personality and intelligence variables did not correlate significantly with HbA1 values. Older patients with shorter duration of diabetes had significantly better glycaemic control (p < 0.05). A significant correlation was observed between HbA1 concentration and onset-age of diabetes (p < 0.001); the patients who had developed diabetes later in life were achieving better control of their blood glucose. In the larger number of subjects in study 2 no significant correlations were evident between HbA1 and personality variables. It is concluded that the predictors of glycaemic control indexed by HbA1 may be distinct from predictors of self-report compliance and that the latter have limited or no value in providing an assessment of quality of glycaemic control. There is no evidence of an effect of personality on glycaemic control as measured by HbA1.

Event related potentials, reaction time, and cognitive performance in idiopathic Parkinson's disease

Sixteen non-demented patients with idiopathic Parkinson's disease (PD) with varying degrees of cognitive impairment and sixteen age-, sex- and education-matched normal controls were examined with (1) an auditory oddball paradigm requiring counting or a motor response in separate determinations, (2) a reaction time task with movement time component and (3) a detailed clinical and neuropsychological test battery. Patients were impaired on a number of neuropsychological tests. They also showed an increased P2 and N2 latency, but no significant increase in P3 latency. Their response initiation times and reaction times during the oddball experiment were not different from controls, whereas movement time was significantly increased. Increased peak latencies, particularly for N2, were moderately associated with Parkinsonian motor impairment in patients and with the Benton Multiple Choice Visual Retention Test in patients and controls. Movement time was associated with P3 latency only in controls and in both groups with the Benton Multiple Choice Visual Retention Test. The observed pattern of results suggests that in non-demented PD patients ERP peak latencies, visuo-spatial task performance and Parkinsonian motor impairment share a significant degree of variance. While impairments in neuropsychological tests and delay in the earlier peaks P2 and N2 do not appear to be sensitive to medication with L-DOPA, normal P3 latencies might indicate good pharmacological symptom control in the absence of dementia.

Cumulative cognitive impairment following recurrent severe hypoglycaemia in adult patients with insulin-treated diabetes mellitus

To examine the hypothesis that episodes of severe hypoglycaemia may cause cumulative cognitive impairment. 100 Type 1 (insulin-dependent) diabetic patients were examined. Their age range was 25-52 years, and the onset of diabetes had occurred after the age of 19 years. Patients with evidence of organic brain disease, including cerebrovascular disease, were excluded. A questionnaire was used to assess the number, frequency and severity of hypoglycaemic episodes experienced during treatment with insulin and the accuracy of this retrospective information was verified from general practice and hospital case-notes. A detailed neuropsychological assessment was undertaken, including tests of pre-morbid and present IQ (Wechsler-Revised), memory and information-processing speed. Significant correlations were observed between the frequency of severe hypoglycaemia and the magnitude of intellectual decline, Performance IQ, inspection time and reaction time (patients with the more frequent hypoglycaemia had poorer performance). Two sub-groups of patients were identified on the basis of their experience of severe hypoglycaemia, and were balanced for pre-morbid IQ, age and duration of diabetes. One sub-group (n = 23) had never experienced severe hypoglycaemia (Group A), whilst the other sub-group (n = 24) had suffered at least five episodes of severe hypoglycaemia (Group B). Group B had greater intellectual impairment than Group A, and Group B also had a significantly slower mean reaction time and higher reaction time variance when compared with Group A.