A Multivariate Interpretation of the Spanish-Language NIH Toolbox Cognition Battery: The Normal Frequency of Low Scores

Abstract

Objective

The current study involved the preparation of multivariate base rates for the Spanish-language NIH Toolbox Cognition Battery (NIHTB-CB) based on the U.S. normative sample, quantifying the normal frequency of low scores among healthy adults.

Method

Participants included 250 healthy Latinx adults (M = 38.8 ± 13.7 years old, range: 19–80; 72.0% women; education: M = 11.5 ± 3.9 years) who completed the full Spanish-language NIHTB-CB, including two tests of crystallized cognition and five tests of fluid cognition. Multivariate base rates quantified the frequency at which participants obtained 1 or more fluid scores ≤25th, ≤16th, ≤9th, ≤5th, and ≤2nd percentile, per age-adjusted or demographically adjusted (age, gender, education) normed scores.

Results

A substantial minority of participants had 1 or more low scores (e.g., 40.4% had 1 or more age-adjusted score ≤16th percentile). The frequencies of low scores increased with fewer years of education and lower crystallized cognitive ability. Higher frequencies of low scores were observed among participants who were born and educated abroad, versus within the USA; monolingual Spanish speakers, versus bilingual Spanish/English speakers; and from households below the national median income, versus households above the national median.

Conclusion

Low scores were common and related to crystallized ability, education, and sociocultural variables. Although using demographically adjusted scores reduced group differences related to sociocultural variables, group differences were not eliminated, indicating that age, gender, and education score adjustments do not fully explain the associations between sociocultural variables and test performances. These stratified base rates may be useful when working with Spanish-speaking patients with diverse sociocultural characteristics.

The Latinx community represents ~18% of the U.S. population (Vespa, Armstrong, & Medina, 2018), with about 75% reporting speaking Spanish at home (Ryan, 2013). Compared with non-Latinx adults, this community is at comparable or greater risk for numerous medical conditions that can adversely affect cognitive functioning (e.g., diabetes, HIV/AIDS, and dementia) (Mayeda, Glymour, Quesenberry, & Whitmer, 2016; National Center for Health Statistics, 2019), potentially attributable to multiple variables including barriers to accessing healthcare and socioeconomic marginality (Vega, Rodriguez, & Gruskin, 2009). These health disparities indicate a population-specific need for neuropsychological services. Few Spanish-language assessment batteries are available for use in clinical practice (Morlett Paredes et al., 2020) and the absence of representative normative data and demographic adjustments for norm-referenced scores can lead to misdiagnosis of cognitive impairment (Cherner et al., 2007; Daugherty, Puente, Fasfous, Hidalgo-Ruzzante, & Pérez-Garcia, 2017). Considering these findings, neuropsychologists require representative normative data and methods for accurately detecting cognitive impairment, ensuring that a diagnosis of cognitive impairment reflects an injury or disease process with minimal influence from demographic characteristics or language of administration.

The National Institutes of Health Toolbox Cognition Battery (NIHTB-CB) is tablet-assisted neuropsychological test battery with English- and Spanish-language versions, separate nationally representative normative samples for each language (Beaumont et al., 2013), and demographic adjustments for multiple characteristics (e.g., age, gender, education, and race/ethnicity) (Casaletto et al., 2015, 2016). The NIHTB-CB has the potential to improve the detection of cognitive weaknesses and impairments in Spanish-speaking patients and research participants, but clinicians require methods to aide in interpreting the battery. Multivariate base rates quantify the frequency at which healthy individuals obtain one or more low or high scores on a cognitive test battery. In research, multivariate base rates have been used to distinguish between severities of traumatic brain injury (Holdnack, Iverson, Silverberg, Tulsky, & Heinemann, 2017a) and to predict cognitive decline (Agelink Van Rentergem et al., 2019). In practice, they can inform neuropsychological test interpretation and they can reduce the likelihood that clinicians will over-interpret a single low score as indicative of cognitive impairment (Binder, Iverson, & Brooks, 2009; Brooks, Iverson, Feldman, & Holdnack, 2009a; Brooks, Iverson, & Holdnack, 2013; Brooks, Iverson, & White, 2007; Karr, Garcia-Barrera, Holdnack, & Iverson, 2017, 2018; Mistridis et al., 2015). Multivariate base rates are available for the English-language NIHTB-CB (Holdnack et al., 2017b), but not yet for the Spanish-language version of the battery.

Researchers have begun developing multivariate base rates for Spanish-language batteries (Olabarrieta-Landa et al., 2019; Rivera et al., 2019). However, these base rates are based on populations throughout Latin America and do not consist of Spanish-speaking adults residing within the USA. Among the few normed batteries available for Spanish-speaking adults (Morlett Paredes et al., 2020), the Spanish-language NIHTB-CB offers the first tablet-assisted neuropsychological test battery for assessing the growing Spanish-speaking U.S. population in research and clinical practice, having undergone a rigorous development process to ensure the reliability and validity of its scores (Fox, Manly, Slotkin, Devin Peipert, & Gershon, 2021; Gershon et al., 2020). The normative sample also has rich sociocultural diversity in terms of national heritage, educational background (e.g., years and country), languages spoken, and socioeconomic status—all of which have a potential relationship with cognitive test performances (Arentoft et al., 2015; Casaletto et al., 2016; Flores et al., 2017). The current study involved the preparation of multivariate base rates of age-adjusted and demographically adjusted scores using the Spanish-language NIHTB-CB for use in research and clinical practice with stratifications based on crystallized ability, years of education, and sociocultural variables. Consistent with prior research, low scores were expected to occur commonly in the Spanish-language NIHTB-CB normative sample, have an inverse relationship with education and intellectual ability, and differ based on sociocultural variables historically related to neuropsychological performance (e.g., country of birth and education, monolingual vs. bilingual, and socioeconomic status).

Method

Participants

The normative sample for the Spanish-language NIHTB-CB (Gershon, 2016) includes 408 adult participants between the ages of 18 and 89 of whom 288 completed all seven tests. Participants were recruited via a planned multisite national sampling approach as part of the same norming study as the English-language NIHTB-CB (Beaumont et al., 2013), with participants who preferred to read in Spanish evaluated in Spanish (Gershon et al., 2020). For this study, participants were excluded if they had incomplete data (n = 120) (i.e., were missing a score on one or more of the NIHTB-CB tests) or reported a pre-existing condition or disorder that is known to affect cognitive functioning (n = 38). The specific self-reported conditions used as exclusion criteria are listed as variables in the NIHTB-CB normative dataset. These conditions included: a pre-existing neurodevelopmental disorder, including attention-deficit/hyperactivity disorder (n = 1), autism spectrum disorder (n = 1), pervasive developmental disorder (n = 1), or a developmental delay (n = 1); a psychiatric or substance use disorder, including a serious emotional disturbance (n = 2), depression or anxiety (n = 29), alcohol abuse (n = 3), or a hospitalization due to emotional problems (n = 4); or a neurological disorder, including epilepsy or seizures (n = 1), Parkinson’s disease (n = 1), or multiple sclerosis (n = 1). Some participants had more than one of these conditions, which is why the sum of these counts exceeds the number of excluded participants. The final sample included 250 participants whose demographic characteristics are reported in Table 1.

Table 1.

Demographic descriptive statistics for the Spanish-language NIHTB-CB normative sample

Demographic Characteristic	Descriptive Statistic
Age (years), M (SD), range	38.8 (13.7), 19–80
Gender, n (%)	—
Men	70 (28.0)
Women	180 (72.0)
Education (years), M (SD), range	11.5 (3.9), 0–20
Education (years), n (%)	—
<12 years	117 (46.8)
12 years	46 (18.4)
13–15 years	39 (15.6)
≥16 years	48 (19.2)
Education (location), n (%)	—
Educated within USA	129 (51.6)
No schooling in USA	118 (47.2)
Pan-Ethnicity, n (%)	—
Latinx	250 (100)
Race, n (%)	—
European	199 (79.6)
Native American	23 (9.2)
African American	12 (4.8)
Multiracial	3 (1.2)
Asian/Pacific Islander	2 (0.8)
Not provided	11 (4.4)
Ethnic Heritage, n (%)	—
Mexican	137 (54.8)
Puerto Rican	11 (4.4)
Cuban	4 (1.6)
Other	95 (38.0)
Not Provided	3 (1.2)
Birthplace, n (%)	—
Born in United States	57 (22.8)
Born Abroad	191 (76.6)
Language, n (%)	—
Monolingual Spanish	189 (75.6)
Bilingual	61 (24.4)
First-language Spanish	233 (93.2)
Primarily Spanish-speaking at home	161 (64.4)
Spanish and English equally at home	55 (22.0)
Household Income, n (%)	—
< $5,000	25 (10.0)
$5,000–$9,999	28 (11.2)
$1,000–$19,999	32 (12.8)
$20,000–$39,999	55 (22.0)
$40,000–$74,999	37 (14.8)
$75,000–$99,999	13 (5.2)
$100,000 or more	8 (3.2)
Do not know	44 (17.6)

Measures

The Spanish-language NIHTB-CB consists of seven tests, two tests of crystallized abilities and five tests of fluid ability (Gershon et al., 2013). The crystallized tests include the Picture Vocabulary test, which involves examinees selecting a picture that corresponds to an aurally presented Spanish word, and the Oral Reading Recognition test, which involves examinees pronouncing Spanish words with the accents removed. The Spanish-language crystallized tests were developed independently of the English-language tests but follow the same administration format (Gershon et al., 2020). Both tests are computerized adaptive, meaning that not all participants are shown the same stimuli. These tests have shown convergent validity with existing and established Spanish-language tests of estimated premorbid intelligence (Fox et al., 2021). They form a crystallized composite score, which was used as an estimate of overall intellectual ability. The frequency of missing data for each crystallized test was as follows: Picture Vocabulary Test (n = 1, 0.3%) and Oral Reading Recognition (n = 10, 2.9%). The reasons for missing data were not known.

The five fluid tests combine to form a fluid composite score. They include tests of visual memory, processing speed, and executive functions. The fluid tests involve minimal use of language aside from the instructions, which were translated from English into Spanish (Gershon et al., 2020). The Picture Sequence Memory test involves the presentation of a series of pictures one at a time; after which the pictures are shuffled, and examinees reorganize them to the order initially presented (Dikmen et al., 2014). The Pattern Comparison Processing Speed test involves examinees deciding, as quickly as possible, whether two presented images are identical or different (Carlozzi et al., 2014). The List Sorting Working Memory test involves the presentation of a sequence of images and corresponding words, which examinees report back in order of size and semantic category (Tulsky et al., 2014). The Flanker Inhibitory Control and Attention test involves examinees seeing a series of arrows, with the middle arrow oriented in either the same or different direction as flanking arrows on each side. The examinee must select a response corresponding to the direction of the middle arrow as quickly as possible (Zelazo et al., 2014). Lastly, the Dimensional Change Card Sort test involves examinees matching an image to target images based on an identified dimension (i.e., shape or color). The dimension requiring matching varies randomly between trials (Zelazo et al., 2014). The frequency of missing data for each fluid test was as follows: Picture Vocabulary Test (n = 1, 0.3%), Oral Reading Recognition (n = 10, 2.9%), Picture Sequence Memory (n = 85; 25.0%), Pattern Comparison Processing Speed (n = 6; 1.8%), List Sorting Working Memory (n = 7, 2.1%), Flanker Inhibitory Control and Attention (n = 1, 0.3%), and Dimensional Change Card Sort (n = 6, 1.8%). As with missing data on crystallized tests, the reasons for missing data on fluid tests were not known.

Statistical analyses

The multivariate base rates were calculated separately for age-adjusted Standard Scores (SS; M = 100, SD = 15) and demographically adjusted T-scores (M = 50, SD = 10) adjusting for age, gender, and education. These normed scores are calculated from raw scores using published formulas consistent with those used by the NIHTB-CB software for automatic scoring in research and clinical practice (Casaletto et al., 2016). Both the age-adjusted and demographically adjusted scores are calculated based on the normative sample of Latinx participants who elected to be evaluated in Spanish, meaning the normed scores gauge performance in comparison to the U.S. Spanish-speaking Latinx population, with adjustments for age only (i.e., for the age-adjusted scores) or adjustments for age, gender, and years of education (i.e., for the demographically adjusted scores). The base rates are calculated as cumulative percentages corresponding to the percent of participants with one or more scores at or below commonly used clinical cutoffs for defining a low score (i.e., ≤25th, ≤16th, ≤9th, ≤5th, and ≤2nd percentiles). The base rates were stratified by crystallized composite (i.e., Low Average: SS ≤ 89/T ≤ 43, Average: SS = 90–99/T = 44–49, Average: SS = 100–109/T = 50–56; High Average: SS ≥ 110/T ≥ 57) and years of education (i.e., <12, 12, 13–15, and ≥16 years). Additional stratifications included on country of birth (i.e., USA vs. Abroad), country of education (i.e., USA vs. Abroad), language (i.e., monolingual vs. bilingual), and annual household income (i.e., <$40,000 vs. ≥$40,000), with incomes categorized as above or below the median Latinx household income for 2010 (DeNavas-Walt, Proctor, & Smith, 2011). For the bilingual stratification, language proficiency was based on self-report, and language of administration was based on preference of the examinee. These sociocultural stratifications were based on variables associated with cognitive test performances based on previous research (Arentoft et al., 2015; Casaletto et al., 2016; Flores et al., 2017).

Results

The multivariate base rates of low scores for the Spanish-language NIHTB-CB for age-adjusted and demographically adjusted norm scores, with stratifications by education and crystallized ability, are provided in Table 2. More than half of the participants completing the five fluid cognition tests on the Spanish-language NIHTB-CB had at least one age-adjusted score at or below the 25th percentile (i.e., 54.8%), and a substantial minority had at least one age-adjusted score at or below the 16th percentile (i.e., 40.4%), 9th percentile (i.e., 25.6%), and 5th percentile (i.e., 18.4%). Fewer participants had two or more low age-adjusted scores, with 31.2% of participants having two or more scores at or below the 25th percentile, 15.2% having two or more scores at or below the 16th percentile, and 7.6% having two or more scores at or below the 9th percentile. This same pattern was observed for the demographically adjusted scores, with a majority having one or more low scores at or below the 25th percentile (i.e., 61.2%) and substantial percentages of participants having one or more scores at or below the 16th percentile (i.e., 46.4%), the 9th percentile (i.e., 25.6%), and the 5th percentile (i.e., 19.2%).

Table 2.

Base Rates of low scores on five Spanish-language NIHTB-CB fluid measures in the normative sample, with stratifications by education and crystallized ability level

	Age	Age norms by yrs. of education					Age norms by crystallized composite				Demo.	Demo. norms by crystallized composite
	norms	<12	12	≥12	13–15	≥16	SS ≤ 89	SS = 90–99	SS = 100–109	SS ≥ 110	norms	T ≤ 43	T = 44–49	T = 50–56	T ≥ 57
Sample Size	250	117	46	133	39	48	67	62	57	64	250	77	45	63	65
≤25th percentile
5 low scores	2.4	5.1	—	—	—	—	9.0	—	—	—	1.6	3.9	—	1.6	—
4 or more	9.6	17.9	2.2	2.3	2.6	2.1	20.9	4.8	5.3	6.3	5.2	9.1	4.4	3.2	3.1
3 or more	18.0	32.5	10.9	5.3	2.6	2.1	35.8	17.7	8.8	7.8	13.6	20.8	15.6	7.9	9.2
2 or more	31.2	45.3	28.3	18.8	17.9	10.4	50.7	30.6	24.6	17.2	30.4	42.9	33.3	22.2	21.5
1 or more	54.8	67.5	54.3	43.6	43.6	33.3	67.2	54.8	54.4	42.2	61.2	70.1	62.2	66.7	44.6
No low scores	45.2	32.5	45.7	56.4	56.4	66.7	32.8	45.2	45.6	57.8	38.8	29.9	37.8	33.3	55.4
≤16th percentile
5 low scores	0.4	0.9	—	—	—	—	1.5	—	—	—	0.8	1.3	—	1.6	—
4 or more	2.4	4.3	2.2	0.8	—	—	7.5	—	1.8	—	2.8	5.2	2.2	3.2	—
3 or more	7.2	13.7	2.2	1.5	—	2.1	17.9	3.2	1.8	4.7	7.2	11.7	11.1	3.2	3.1
2 or more	15.2	28.2	8.7	3.8	—	2.1	32.8	14.5	5.3	6.3	18.0	23.4	20.0	12.7	15.4
1 or more	40.4	53.8	34.8	28.6	30.8	20.8	59.7	33.9	31.6	34.4	46.4	59.7	44.4	42.9	35.4
No low scores	59.6	46.2	65.2	71.4	69.2	79.2	40.3	66.1	68.4	65.6	53.6	40.3	55.6	57.1	64.6
≤9th percentile
5 low scores	0.4	0.9	—	—	—	—	1.5	—	—	—	—	—	—	—	—
4 or more	1.6	3.4	—	—	—	—	4.5	—	1.8	—	0.8	1.3	—	1.6	—
3 or more	2.4	5.1	—	—	—	—	7.5	—	1.8	—	3.2	5.2	2.2	3.2	1.5
2 or more	7.6	14.5	2.2	1.5	–	2.1	17.9	3.2	3.5	4.7	7.6	9.1	8.9	7.9	4.6
1 or more	25.6	35.9	21.7	16.5	15.4	12.5	43.3	19.4	15.8	21.9	25.6	36.4	20.0	20.6	21.5
No low scores	74.4	64.1	78.3	83.5	84.6	87.5	56.7	80.6	84.2	78.1	74.4	63.6	80.0	79.4	78.5
≤5th percentile
5 low scores	0.4	0.9	—	—	—	—	—	—	—	—	—	—	—	—	—
4 low scores	0.4	0.9	—	—	—	—	1.5	—	—	—	—	—	—	—	—
3 or more	1.6	3.4	—	—	—	—	4.5	—	1.8	—	0.8	1.3	—	1.6	—
2 or more	3.6	6.8	—	0.8	—	2.1	7.5	1.6	3.5	1.6	4.4	5.2	2.2	6.3	3.1
1 or more	18.4	28.2	13.0	9.8	7.7	8.3	37.3	14.5	8.8	10.9	19.2	28.6	13.3	14.3	16.9
No low scores	81.6	71.8	87.0	90.2	92.3	91.7	62.7	85.5	91.2	89.1	80.8	71.4	86.7	85.7	83.1
≤2nd percentile
3 or more	0.4	0.9	—	—	—	—	1.5	—	—	—	0.4	1.3	—	—	—
2 or more	2.0	3.4	—	0.8	—	2.1	4.5	—	1.8	1.6	2.0	2.6	2.2	3.2	—
1 or more	10.0	17.1	2.2	3.8	2.6	6.3	19.4	9.7	5.3	4.7	12.0	20.8	8.9	9.5	6.2
No low scores	90.0	82.9	97.8	96.2	97.4	93.8	80.6	90.3	94.7	95.3	88.0	79.2	91.1	90.5	93.8

Note. Demo. = Demographically adjusted; NIHTB-CB = National Institutes of Health Toolbox for the Assessment of Neurological and Behavioral Function Cognition Battery; SS = Standard Score; T = T-score; Yrs. = Years. Values represent cumulative percentages except for the rows labeled “No low scores,” which provide the percentage of the normative sample with no scores falling below the low score cut-offs. Age norms are provided as age adjusted SSs (M = 100, SD = 15) and demographically adjusted norms are provided as T-scores (M = 50, SD = 10) adjusted for age, sex, and education. The NIHTB-CB includes five tests of fluid abilities: Flanker Inhibitory Control and Attention Test, Picture Sequence Memory Test, List Sorting Working Memory Test, Dimensional Change Card Sort Test, and Pattern Comparison Processing Speed Test. Two tests measure crystallized abilities, from which the crystallized composite is calculated: Picture Vocabulary Test and Oral Reading Recognition Test.

Participants with more education or higher estimated crystallized ability had fewer low test scores. Roughly a quarter of participants (i.e., 28.2%) with less than 12 years of education obtained two or more age-adjusted scores at or below the 16th percentile, whereas just 2.1% of participants with 16 or more years of education had this same pattern of low scores. Participants with below average crystallized ability levels (i.e., either SS ≤ 89 or T ≤ 43) had the highest frequency of low scores, and people with higher performances on measures of crystallized ability are less likely to obtain one or more low fluid cognitive test scores. The rates of obtaining one or more demographically adjusted scores at or below the 16th percentile for participants with crystallized composites of T ≤ 43, T = 44–49, T = 50–56, and T ≥ 57 were 59.7%, 44.4%, 42.9%, and 35.4%, respectively. In contrast, using a lower criterion for defining a low score (i.e., the 9th percentile), those in the average to above average crystallized ability stratifications had roughly similar frequencies of low scores. The rates of one or more demographically adjusted score at or below the 9th percentile were 36.4%, 20.0%, 20.6%, and 21.5% for the T ≤ 43, T = 44–49, T = 50–56, and T ≥ 57 stratifications, respectively.

Additional demographic stratifications are provided in Table 3, including country of birth, country of education, monolingual versus bilingual status, and annual household income. Participants who were born abroad had more low scores than participants born in the USA. For example, 49.7% of participants born abroad obtained one or more demographically adjusted scores at or below the 16th percentile compared with 35.1% of those born in the USA. Participants educated abroad had more low scores than participants educated in the USA. For example, 51.7% of those educated abroad obtained one or more age-adjusted scores at or below the 16th percentile compared with 30.2% of those educated in the USA. Participants who were monolingual Spanish speakers obtained more age-adjusted low scores than participants who were bilingual, but preferred to communicate in Spanish (e.g., 45.0% of monolingual participants vs. 26.2% of bilingual participants obtained one or more scores at or below the 16th percentile). However, the base rates for these groups were more comparable when using demographically adjusted norms (e.g., 47.1% of monolingual participants vs. 44.3% of bilingual participants obtained one or more scores at or below the 16th percentile). In terms of annual household income, participants from households making <$40,000 per year obtained more low scores than participants from households making ≥$40,000 per year. The magnitude of difference between the groups in each of these stratifications was smaller when using demographically adjusted norms compared with age-adjusted norms (see Fig. 1).

Table 3.

Base Rates of low scores on five Spanish-language NIHTB-CB fluid measures in the normative sample, with stratifications by demographic and language characteristics

	Age Norms (SS)								Demographically Adjusted Norms (T)
	Birthplace		Education		Language		Income		Birthplace		Education		Language		Income
	Abroad	USA	No School in USA	School in USA	Mono-lingual	Bilingual	<$40 k	≥$40 k	Abroad	U.S.	No School in U.S.	School in U.S.	Monolingual	Bilingual	<$40 k	≥$40 k
Sample Size	191	57	118	129	189	61	140	58	191	57	118	129	189	61	140	58
≤25th percentile
5 low scores	3.1	–	3.4	0.8	3.2	–	1.4	–	2.1	–	2.5	0.8	2.1	–	1.4	–
4 or more	11.0	5.3	14.4	4.7	12.2	1.6	5.0	–	5.8	3.5	6.8	3.1	6.3	1.6	2.9	–
3 or more	22.0	5.3	28.0	8.5	22.8	3.3	13.6	1.7	16.2	5.3	19.5	7.8	16.9	3.3	10.0	1.7
2 or more	37.2	10.5	44.1	18.6	37.0	13.1	30.7	3.4	34.0	17.5	36.4	24.0	33.3	21.3	30.7	8.6
1 or more	61.8	31.6	67.8	42.6	59.3	41.0	59.3	24.1	62.8	56.1	63.6	58.1	62.4	57.4	67.1	37.9
No low scores	38.2	68.4	32.2	57.4	40.7	59.0	40.7	75.9	37.2	43.9	36.4	41.9	37.6	42.6	32.9	62.1
≤16th percentile
5 low scores	0.5	—	—	0.8	0.5	—	0.7	—	1.0	—	0.8	0.8	1.1	—	0.7	—
4 or more	3.1	—	2.5	1.6	3.2	—	1.4	—	3.7	—	3.4	1.6	3.7	—	1.4	—
3 or more	8.9	1.8	9.3	4.7	9.5	—	3.6	—	8.4	3.5	9.3	4.7	9.0	1.6	6.4	—
2 or more	18.8	3.5	22.9	7.8	19.6	1.6	13.6	—	21.5	7.0	24.6	11.6	21.2	8.2	18.6	5.2
1 or more	46.1	22.8	51.7	30.2	45.0	26.2	46.4	8.6	49.7	35.1	52.5	40.3	47.1	44.3	50.0	22.4
No low scores	53.9	77.2	48.3	69.8	55.0	73.8	53.6	91.4	50.3	64.9	47.5	59.7	52.9	55.7	50.0	77.6
≤9th percentile
5 low scores	0.5	—	—	0.8	0.5	—	0.7	—	–	—	—		—	—	—	—
4 or more	2.1	—	1.7	0.8	2.1	—	1.4	—	1.0	—	0.8	0.8	1.1	—	0.7	—
3 or more	3.1	—	2.5	1.6	3.2	—	1.4	—	4.2	–	4.2	1.6	4.2	—	2.1	—
2 or more	9.9	—	10.2	4.7	9.5	1.6	5.0	—	9.4	1.8	10.2	4.7	9.5	1.6	6.4	—
1 or more	30.4	10.5	33.9	17.8	30.2	11.5	27.9	3.4	28.3	17.5	30.5	20.9	27.5	19.7	25.0	13.8
No low scores	69.6	89.5	66.1	82.2	69.8	88.5	72.1	96.6	71.7	82.5	69.5	79.1	72.5	80.3	75.0	86.2
≤5th percentile
5 low scores	0.5	—	—	0.8	0.5	—	0.7	—	—	—	—		—	—	—	—
4 low scores	0.5	—	—	0.8	0.5	—	0.7	—	—	—	—		—	—	—	—
3 or more	2.1	—	1.7	0.8	2.1	—	1.4	—	1.0	—	0.8	0.8	1.1	—	0.7	—
2 or more	4.7	—	5.1	1.6	4.8	—	2.9	—	5.8	—	5.9	2.3	5.8	—	2.9	—
1 or more	21.5	8.8	25.4	11.6	22.8	4.9	17.9	3.4	22.0	10.5	25.4	13.2	22.8	8.2	18.6	6.9
No low scores	78.5	91.2	74.6	88.4	77.2	95.1	82.1	96.6	78.0	89.5	74.6	86.8	77.2	91.8	81.4	93.1
≤2nd percentile
3 or more	0.5	—	—	—	0.5	—	—	—	0.5	—	—	0.8	0.5	—	0.7	—
2 or more	2.6	—	2.5	0.8	2.6	—	1.4	—	2.6	—	2.5	0.8	2.6	—	1.4	—
1 or more	12.0	3.5	12.7	7.0	11.6	4.9	8.6	1.7	14.1	5.3	16.1	7.8	14.3	4.9	12.1	3.4
No low scores	88.0	96.5	87.3	93.0	88.4	95.1	91.4	98.3	85.9	94.7	83.9	92.2	85.7	95.1	87.9	96.6

Note. NIHTB-CB = National Institutes of Health Toolbox for the Assessment of Neurological and Behavioral Function Cognition Battery; SS = Standard Score; T = T-score; Values represent cumulative percentages except for the rows labeled “No low scores,” which provide the percentage of the normative sample with no scores falling below the low score cut-offs. Age norms are provided as age adjusted SSs (M = 100, SD = 15) and demographically adjusted norms are provided as T-scores (M = 50, SD = 10) adjusted for age, sex, and education. Income is self-reported annual household income the year prior to the assessment. The NIHTB-CB includes five tests of fluid abilities: Flanker Inhibitory Control and Attention Test, Picture Sequence Memory Test, List Sorting Working Memory Test, Dimensional Change Card Sort Test, and Pattern Comparison Processing Speed Test. Two tests measure crystallized abilities, from which the crystallized composite is calculated: Picture Vocabulary Test and Oral Reading Recognition Test.

Fig. 1.

Percentage of healthy adults with one or more fluid cognitive test scores at or below the 16th percentile on the Spanish-language NIH Toolbox Cognition Battery, with stratifications based on crystallized ability, country of origin and education, monolingual versus bilingual status, and annual household income.

Discussion

This is the first study to examine the prevalence of low scores on the Spanish-language NIHTB-CB among healthy adults. Consistent with prior multivariate base rates research, participants commonly obtained one or more low scores on the Spanish-language NIHTB-CB (Binder et al., 2009; Brooks et al., 2013; Brooks, Iverson, Lanting, Horton, & Reynolds, 2012; Brooks, Iverson, & White, 2009b; Holdnack, Tulsky, et al., 2017b; Karr et al., 2017, 2018). Moreover, people with less education and lower intellectual ability, as measured by the crystallized composite score, were more likely to obtain low fluid cognitive test scores—consistent with prior studies using other test batteries (Brooks et al., 2013; Karr et al., 2017, 2018). The proportions of participants with low scores on the Spanish-language NIHTB-CB were fairly similar to the proportions of participants obtaining low scores on the English-language NIHTB-CB (see Table 3 in Holdnack, Tulsky, et al., 2017b). There were modest differences in the base rates, and in some subgroups, the participants completing the battery in Spanish had fewer low scores than those completing the battery in English. These modest differences likely relate, in part, to the fact that any participant with a pre-existing condition that might be associated with lower fluid cognitive test scores was excluded from this study, but not from the study on the English-language battery (Holdnack, Tulsky, et al., 2017b).

A notable nonintuitive finding pertained to the frequency of low scores among participants with high average crystallized composite scores: these participants had a comparable or greater frequency of low scores than participants with average crystallized composites. This trend was not apparent for counts of age-adjusted or demographically adjusted scores ≤25th percentile but was apparent for counts of age-adjusted scores ≤16th, ≤9th, and ≤5th and demographically adjusted scores ≤9th, ≤5th, and ≤2nd percentile. One reason for this nonintuitive finding may be the small sample sizes for the stratifications (n range: 57–67), which resulted in very small counts of participants obtaining one or more scores at or below the lower percentile cutoffs. The nonintuitive trend could be an artifact of insufficient sampling to get a proper population estimate of low scores for these crystallized ability stratifications and may not replicate if larger samples were collected for these stratifications.

Another reason for this nonintuitive trend may be the association between sociocultural variables and cognitive test performances. Spanish speakers born and educated abroad, who speak Spanish as their first language, and primarily speak Spanish at home have higher crystallized test scores, whereas Spanish speakers born and educated within the USA, who speak Spanish and English equally at home have higher fluid test scores (Casaletto et al., 2016; Flores et al., 2017). Based on these relationships, participants with higher Spanish-language education who speak Spanish daily will obtain higher crystallized scores but lower fluid scores, which could account, in part, for the counterintuitive relationship observed in Table 2.

Participants with low average crystallized composites had the highest frequency of low scores across all percentile cutoffs, indicating a correspondence between crystallized and fluid cognitive test performances at this ability level. The crystallized composite may be an acceptable estimate of premorbid functioning among participants with low average crystallized ability. However, the nonlinear relationship between crystallized composite and low fluid test scores observed for average to high average crystallized ability stratifications may result in part from individual differences in sociocultural experiences. As such, the crystallized composite may not provide the best estimate of premorbid ability among some Spanish-speaking examinees with average or high average crystallized ability, and sociocultural variables should be considered in combination with this estimate.

A new insight from this study pertains to the demographic stratifications, including country of birth, country of education, monolingual versus bilingual status, and household income. A substantially higher percentage of low scores were obtained by participants who were born outside of the USA, were educated outside of the USA, were monolingual Spanish speakers, and were from households with annual incomes below the national median (i.e., <$40,000). These stratifications emphasize the relationship between different sociocultural variables and performances on neuropsychological tests. These relationships may be associated with differences in acculturative experiences (e.g., between participant born and educated abroad vs. within the USA), as described by prior researchers (Arnold, Montgomery, Castañeda, & Longoria, 1994; Boone, Victor, Wen, Razani, & Pontón, 2007; Coffey, Marmol, Schock, & Adams, 2005). They may have other explanations, such as a bilingual advantage explaining higher fluid performances by bilingual participants (Bialystok, Klein, Craik, & Viswanathan, 2004); however, prior research syntheses have challenged the notion of a bilingual advantage (Lehtonen et al., 2018). Spanish-English bilingualism and acculturative experiences within the USA are likely closely related (e.g., bilinguals may have resided longer within the USA, been educated within the USA, and consume English-language American media to a greater degree), and differences in acculturation may more likely explain differences in low score frequencies in the current sample. This conclusion is limited though, in that a direct objective measure of acculturation was not administered, and the relationship between acculturation and cognitive test performances requires further research in general (Tan, Burgess, & Green, 2020).

The sociocultural stratifications show the potential influence of these variables even when participants are evaluated in their preferred language and scores are adjusted for age, gender, and years of education. These stratifications may be useful when working with Spanish-speaking patients with diverse sociocultural characteristics and experiences. For example, for the full sample, 18.0% of participants obtained two or more demographically adjusted scores at or below the 16th percentile. When considering country of birth, a similar percentage of patients who were born abroad had this performance pattern (i.e., 21.5%); however, only 7.0% of those born in the USA obtained this same pattern of scores. In terms of annual household income, 50.0% of those below the national median household income obtained two of more demographically adjusted scores at or below the 16th percentile, whereas 22.4% of those above the median obtained this number of low scores. These stratifications provide meaningful information about how common a pattern of low performances may be in different subgroups of Spanish-speaking adults.

These multivariate base rates can be immediately translated into clinical practice for use with patients who would prefer to be evaluated in Spanish. Normative classifications are provided in Tables 4 and 5 for ease of use in clinical practice, allowing clinicians to determine whether a count of low scores is Broadly Normal (i.e., occurs in >25% of the normative sample), Below Average (i.e., occurs in <25% of the normative sample), Uncommon (i.e., occurs in <10% of the normative sample), or Very Uncommon (i.e., occurs in <3% of the normative sample). These multivariate base rates offer an additional resource for clinicians interpreting test performances on the Spanish-language NIHTB-CB. They are best used in combination with medical, psychological, and contextual information to inform clinical judgement. If a patient presents with a clinical condition that might adversely affect fluid cognitive functioning (e.g., frontal lobe injury, anterior cerebral artery stroke), the presence of a low score should be interpreted in the context of predicted deficits (e.g., area of brain damage per neuroimaging), corresponding subjective cognitive complaints, and degree of functional disability.

Table 4.

Classification ranges for number of low Spanish-language NIHTB-CB fluid subtest scores, with stratifications by education and crystallized ability level

	Age	Age norms by yrs. of education					Age norms by crystallized composite				Demo.	Demo. norms by crystallized composite
	norms	<12	12	≥12	13–15	≥16	SS ≤ 89	SS = 90–99	SS = 100–109	SS ≥ 110	norms	T ≤ 43	T = 44–49	T = 50–56	T ≥ 57
Sample Size	250	117	46	133	39	48	67	62	57	64	250	77	45	63	65
≤25th percentile
Broadly Normal	0–2	0–3	0–2	0–1	0–1	0–1	0–3	0–2	0–1	0–1	0–2	0–2	0–2	0–1	0–1
Below Average	3	4	3	2	2	2	4	3	2	2	3	3	3	2	2
Uncommon	4	5	—	3	—	—	5	4	3–4	3–4	4	4	4	3–4	3–4
Very Uncommon	5	—	4–5	4–5	3–5	3–5	—	5	5	5	5	5	5	5	5
≤16th percentile
Broadly Normal	0–1	0–2	0–1	0–1	0–1	0–1	0–2	0–1	0–1	0–1	0–1	0–1	0–1	0–1	0–1
Below Average	2	3	—	—	—	—	3	2	—	—	2	2–3	2–3	2	2
Uncommon	3	4	2	2	—	—	4	3	2	2–3	3	4	—	3–4	3
Very Uncommon	4–5	5	3–5	3–5	2–5	2–5	5	4–5	3–5	4–5	4–5	5	4–5	5	4–5
≤9th percentile
Broadly Normal	0–1	0–1	0	0	0	0	0–1	0	0	0	0–1	0–1	0	0	0
Below Average	—	2	1	1	1	1	2	1	—	1	—	—	1	1	1
Uncommon	2	3–4	—	—	—	—	3–4	2	1–2	2	2–3	2–3	2	2–3	2
Very Uncommon	3–5	5	2–5	2–5	2–5	2–5	5	3–5	3–5	3–5	4–5	4–5	3–5	4–5	3–5
≤5th percentile
Broadly Normal	0	0–1	0	0	0	0	0–1	0	0	0	0	0–1	0	0	0
Below Average	1	–	1	—	—	—	—	1	—	1	1	—	1	1	1
Uncommon	2	2–3	—	1	1	1	2–3	—	1–2	—	2	2	—	2	2
Very Uncommon	3–5	4–5	2–5	2–5	2–5	2–5	4–5	2–5	3–5	2–5	3–5	3–5	2–5	3–5	3–5
≤2nd percentile
Broadly Normal	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Below Average	1	1	—	—	—	—	1	—	—	—	1	1	—	—	—
Uncommon	—	—	—	1	—	1	2	1	1	1	—	—	1	1–2	1
Very Uncommon	2–5	2–5	1–5	2–5	1–5	2–5	3–5	2–5	2–5	2–5	2–5	2–5	2–5	3–5	2–5

Note. NIHTB-CB = National Institutes of Health Toolbox for the Assessment of Neurological and Behavioral Function Cognition Battery; SS=Standard Score; T = T-score; Yrs. = Years. Age norms are provided as age adjusted SSs (M = 100, SD = 15) and demographically adjusted norms are provided as T scores (M = 50, SD = 10) adjusted for age, sex, and education. The NIHTB-CB includes five tests of fluid abilities: Flanker Inhibitory Control and Attention Test, Picture Sequence Memory Test, List Sorting Working Memory Test, Dimensional Change Card Sort Test, and Pattern Comparison Processing Speed Test. Two tests measure crystallized abilities, from which the crystallized composite is calculated: Picture Vocabulary Test and Oral Reading Recognition Test. The classification ranges refer to the number of low scores obtained to be considered Broadly Normal (i.e., the number of low scores obtained by the top 75% of the normative sample), Below Average (i.e., the number of low scores obtained by less than 25% of the normative sample), Uncommon (i.e., the number of low scores obtained by less than 10% of the normative sample), and Very Uncommon (i.e., the number of low scores obtained by less than 3% of the normative sample).

Table 5.

Classification ranges for number of low Spanish-language NIHTB-CB fluid subtest scores, with stratifications by demographic and language characteristics

	Age Norms (SS)								Demographically Adjusted Norms (T)
	Birthplace		Education		Language		Income		Birthplace		Education		Language		Income
	Abroad	USA	No School in USA	School in USA	Monolingual	Bilingual	<$40 k	≥$40 k	Abroad	USA	No School in USA	School in USA	Monolingual	Bilingual	<$40 k	≥$40 k
Sample Size	191	57	118	129	189	61	140	58	191	57	118	129	189	61	140	58
≤25th percentile
Broadly Normal	0–2	0–1	0–3	0–1	0–2	0–1	0–2	0	0–2	0–1	0–2	0–1	0–2	0–1	0–2	0–1
Below Average	3–4	2	4	2	3–4	2	3	1	3	2	3	2	3	2	3	–
Uncommon	5	3–4	5	3–4	5	3	4	2	4	3–4	4	3–4	4	3	–	2
Very Uncommon	—	5	—	5	—	4–5	5	3–5	5	5	5	5	5	4–5	4–5	3–5
≤16th percentile
Broadly Normal	0–1	0	0–1	0–1	0–1	0–1	0–1	0	0–1	0–1	0–1	0–1	0–1	0–1	0–1	0
Below Average	2	1	2	—	2	—	2	—	2	—	2	2	2	—	2	1
Uncommon	3–4	2	3	2–3	3–4	–	3	1	3–4	2–3	3–4	3	3–4	2	3	2
Very Uncommon	5	3–5	4–5	4–5	5	2–5	4–5	2–5	5	4–5	5	4–5	5	3–5	4–5	3–5
≤9th percentile
Broadly Normal	0–1	0	0–1	0	0–1	0	0–1	0	0–1	0	0–1	0	0–1	0	0–1	0
Below Average	—	1	2	1	—	1	—	—	—	1	2	1	—	1	—	1
Uncommon	2–3	—	—	2	2–3	—	2	1	2–3	—	3	2	2–3	—	2	–
Very Uncommon	4–5	4–5	3–5	3–5	4–5	2–5	3–5	2–5	4–5	2–5	4–5	3–5	4–5	2–5	3–5	2–5
≤5th percentile
Broadly Normal	0	0	0–1	0	0	0	0	0	0	0	0–1	0	0	0	0	0
Below Average	1	—	—	1	1	—	1	—	1	1	—	1	1	—	1	—
Uncommon	2	1	2	—	2	1	—	1	2	2	2	—	2	1	—	1
Very Uncommon	3–5	4–5	3–5	2–5	3–5	2–5	2–5	2–5	3–5	3–5	3–5	2–5	3–5	2–5	2–5	2–5
≤2nd percentile
Broadly Normal	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Below Average	1	—	1	—	1	—	—	—	1	1	1	—	1	—	1	—
Uncommon	—	1	—	1	—	1	1	—	—	—	—	1	—	1	—	1
Very Uncommon	2–5	4–5	2–5	2–5	2–5	2–5	2–5	1–5	2–5	2–5	2–5	2–5	2–5	2–5	2–5	2–5

Note. NIHTB-CB = National Institutes of Health Toolbox for the Assessment of Neurological and Behavioral Function Cognition Battery; SS = Standard Score; T = T-score; Age norms are provided as age adjusted SSs (M = 100, SD = 15) and demographically adjusted norms are provided as T scores (M = 50, SD = 10) adjusted for age, sex, and education. The NIHTB-CB includes five tests of fluid abilities: Flanker Inhibitory Control and Attention Test, Picture Sequence Memory Test, List Sorting Working Memory Test, Dimensional Change Card Sort Test, and Pattern Comparison Processing Speed Test. Two tests measure crystallized abilities, from which the crystallized composite is calculated: Picture Vocabulary Test and Oral Reading Recognition Test. The classification ranges refer to the number of low scores obtained to be considered Broadly Normal (i.e., the number of low scores obtained by the top 75% of the normative sample), Below Average (i.e., the number of low scores obtained by less than 25% of the normative sample), Uncommon (i.e., the number of low scores obtained by less than 10% of the normative sample), and Very Uncommon (i.e., the number of low scores obtained by less than 3% of the normative sample).

As an example of their use, take a 75-year-old Latinx grandmother, born in Mexico where she completed 6 years of education. She is a monolingual Spanish speaker and emigrated to the USA to live with her children and grandchildren who established roots in southern California. Her grandchildren have expressed concerns about her forgetfulness, noting that she loses items around the home and occasionally leaves the stove on after cooking. She does not feel concerned about her memory and describes her occasional lapses as normal for her age. She has mild hypertension and osteoporosis, and no other known medical conditions. No neuroimaging had been conducted at the time of the assessment. She reported no family history of neurological or neurodegenerative conditions. In terms of psychological health, she described mild symptoms of depression.

The patient completed the Spanish-language NIHTB-CB and obtained a demographically adjusted crystallized composite score of T = 48 (42nd percentile) and the following fluid cognition scores: Picture Sequence Memory (T = 35, 7th percentile), Pattern Comparison Processing Speed (T = 47, 38th percentile), List Sorting Working Memory (T = 39, 14th percentile), Flanker Inhibitory Control and Attention (T = 44, 27th percentile), and Dimensional Change Card Sort (T = 45, 31st percentile). Of note, two scores involving memory performance fall below the 16th percentile, which aligns with the concerns of her family members. She has one fluid score at or below the 9th percentile and two fluid scores at or below the 16th percentile; which, respectively, occur in 25.6% and 18.0% of the normative sample (see Table 2), and would be considered Broadly Normal and Below Average performance patterns (see Table 4). When considering her estimated premorbid function (per her crystallized composite score), these performance patterns are both Below Average, occurring among 20.0% of participants with a crystallized composite of T = 44–49 (see Table 2). Considering sociocultural variables, 21.5% of participants born abroad obtained two or more scores at or below the 16th percentile and 28.3% obtained one or more scores at or below the 9th percentile; 24.6% of participants educated abroad speakers obtained two or more scores at or below the 16th percentile and 30.5% obtained one or more scores at or below the 9th percentile; and 21.2% of monolingual Spanish speakers obtained two or more scores at or below the 16th percentile and 27.5% obtained one or more scores at or below the 9th percentile (see Table 3). For all stratifications considered, it was Broadly Normal to obtain a single score at or below the 9th percentile and slightly Below Average to obtain two or more scores at or below the 16th percentile (see Table 5). The presence of a single score at the 7th percentile, or two scores below the 16th percentile, may be concerning for a clinician, who may consider a diagnosis of mild neurocognitive disorder. However, the multivariate base rates of low scores demonstrate that these performance patterns are not uncommon in the normative sample. It would be reasonable to conclude that this woman has either broadly normal cognitive functioning or modest weaknesses, as measured by the NIHTB-CB, at the present time. Her pattern of performance is not unusually low, however, in a manner that might reflect mild cognitive impairment. She might benefit from a follow-up evaluation, in the future, to monitor her cognitive functioning.

Take another example of a 45-year-old monolingual Spanish-speaking Latino man who emigrated to the USA from Mexico 7 years prior to his evaluation. He has 5 years of formal education in Mexico and worked primarily in physical labor jobs his entire life. He recently experienced a moderate traumatic brain injury due to a workplace accident. A day of injury computed tomography scan was positive for a hematoma in his left frontal lobe, which was managed without neurosurgical intervention. He was discharged from the hospital the next day. After a few months, he was seen by a physiatrist and told her that he felt inattentive, slowed down, and disorganized since his injury, describing problems at work and conflict at home. He was referred for neuropsychological assessment.

This patient reported mild difficulties on Spanish-language questionnaires of anxiety and depression symptoms. He was administered the Spanish-language NIHTB-CB and obtained the following scores: a demographically adjusted crystallized composite score of T = 40 (16th percentile) and the following fluid cognition scores: Picture Sequence Memory (T = 47, 38th percentile), Pattern Comparison Processing Speed (T = 38, 12th percentile), List Sorting Working Memory (T = 42, 21st percentile), Flanker Inhibitory Control and Attention (T = 37, 10th percentile), and Dimensional Change Card Sort (T = 40, 16th percentile). Based on his estimated premorbid functioning, his performances may be considered normal. His crystallized composite is Low Average and his fluid scores include three Low Average scores and two Average scores. He has no scores that would be considered borderline or extremely low. However, he obtains three scores at or below the 16th percentile, which occurs in just 11.7% of the normative sample at his ability level stratification (see Table 2), which would be considered an Uncommon performance pattern (see Table 4). Based on sociocultural stratifications, his performance pattern occurs in 8.4% of Spanish-speaking adults born abroad, 9.3% without education within the USA, and 9.0% of monolingual Spanish speakers (see Table 3), all of which would be interpreted as Uncommon performance profiles (see Table 5). Considering his recent injury and uncommon performances based on his estimated premorbid ability and sociocultural variables, it would be reasonable to conclude that he is continuing to experience objectively measured cognitive difficulties due to traumatic brain injury.

The multivariate base rates can inform interpretation of the Spanish-language NIHTB-CB, but the study has limitations that should be considered when translating these findings into clinical practice. The overall sample size for preparing the multivariate base rates was relatively small, which made sample sizes for various stratifications small as well. The sample was also predominantly women and gender has been associated with NIHTB-CB test performances (Casaletto et al., 2016). Gender is considered in the calculation of the demographically adjusted norms, but not the age-adjusted norms, making the age-adjusted norms less appropriate for use with men. In terms of racial and ethnic diversity, the sample was pan-ethnically Latinx; but most of the sample identified racially as European, and no demographic adjustments for race were used in the normative score calculations. The predominantly Mexican-American sample may limit the generalizability of the findings to patients of different national and regional origins, such as Puerto Ricans, Cubans, and individuals of other national, territorial, or regional heritages (Marquine et al., 2018; Rivera Mindt et al., 2020), although further research is needed to clarify whether different communities within the pan-ethnic Latinx population vary substantially enough in cognitive test performances to rationalize more specific stratifications in normative data. There was also insufficient sample size to provide a stratification for education less than 6 years, which may serve as a critical educational cutoff in Latin America (Welti, 2010), akin to 12 years in the USA. In the total sample, 4.0% (n = 10) had three or fewer years of education and 12.4% (n = 31) had six or fewer years of education. The lowest education stratification (i.e., <12 years) had a mean education of 8.3 years (SD = 2.9; range: 0–11), which may not be representative of participants with very minimal formal education. Participants with no or few years of education may have limited reading ability or be entirely unable to read, which has known effects on cognitive and neurological functioning (Ardila et al., 2010). In these examinees, education stratifications may be less useful for estimating current or premorbid levels of functioning, and emphasis may need to be placed on functional scales as opposed to cognitive test scores (Noroozian, Shakiba, & Iran-nejad, 2014). Although limited, these base rates offer a new approach to interpreting Spanish-language NIHTB-CB data to serve patients traditionally underserved in neuropsychological settings.

Frequent challenges for clinical neuropsychologists include inadequate normative data, culturally biased tests, and a lack of translated tests into different languages (Rabin, Paolillo, & Barr, 2016). Addressing these challenges, researchers have begun to develop new normative datasets for traditional neuropsychological measures, including Spanish-language norming initiatives in Spain (Peña-Casanova et al., 2009), Latin America (Guàrdia-Olmos, Peró-Cebollero, Rivera, & Arango-Lasprilla, 2015), and the USA-Mexico border region (Cherner et al., 2021), with some producing multivariate base rates (Olabarrieta-Landa et al., 2019; Rivera et al., 2019, 2020). The availability of Spanish-language norms from different world regions aligns with the tremendous within-group diversity of Spanish-speakers, and the necessity of normative data that matches the linguistic and cultural backgrounds of patients seen for neuropsychological assessments. Within the NIHTB-CB norming study, Latinx adults elected to complete the battery in their preferred language, and compared with those completing the English-language battery, those choosing to complete the battery in Spanish had unique acculturative experiences (e.g., country of origin and education, languages spoken) and economic experiences, which—as shown in the base rate stratifications—were related to test performances. The multivariate base rates provide a resource to work with this diverse community in a manner that more precisely identifies unusual patterns of neuropsychological test performances that may be indicative of cognitive impairment.

Funding

This research was supported by the National Academy of Neuropsychology Clinical Research Grant (2020-21), Advanced Psychometric Interpretation of the Spanish language NIH Toolbox Cognition Battery: Developing Diagnostic Algorithms for Cognitive Impairment and Interpreting Reliable Change (Principal Investigator: J. E. Karr). This research was also supported, in part, by the National Institute on Aging of the National Institutes of Health (under Award R01AG050720), Study of Aging Latinas/os for Understanding Dementia in HIV (Principal Investigator: M. Rivera Mindt).

Conflict of interest

Dr. Justin E. Karr and Dr. Monica Rivera Mindt have no disclosures. Dr. Grant L. Iverson has received past research funding from several test publishing companies, including ImPACT Applications, Inc., CNS Vital Signs, and Psychological Assessment Resources (PAR, Inc.). He receives royalties from one neuropsychological test (WCST-64). He acknowledges unrestricted philanthropic support from ImPACT Applications, Inc., the Mooney-Reed Charitable Foundation, and the Spaulding Research Institute.