Copenhagen Cross-Linguistic Naming Test (C-CLNT)

Abstract

The Copenhagen Cross-Linguistic Naming Test (C-CLNT; Nielsen et al., 2023) is a standardized neuropsychological instrument developed to assess visual confrontation naming and detect naming impairment (anomia) across culturally, linguistically, and educationally diverse older populations. Originating from the Danish Dementia Research Centre at Copenhagen University Hospital–Rigshospitalet, the tool addresses the profound cross-cultural biases inherent in legacy instruments, such as the Boston Naming Test (BNT). The C-CLNT is derived from the multilingual MULTIMAP database (Gisbert-Muñoz et al., 2021), a standardized repository of 218 color drawings of concrete objects and actions calibrated across multiple European and non-European languages. The standard C-CLNT protocol comprises 30 items (20 target objects and 10 target actions), alongside a validated abbreviated 20-item version focusing exclusively on object naming. Administration adheres to a structured latency-based protocol (5 seconds per item) incorporating hierarchical semantic and phonemic cueing pathways. Psychometric evaluations in memory clinic cohorts demonstrate satisfactory internal consistency (Cronbach’s alpha = .67) and robust construct validity, evidenced by moderate-to-strong correlations with the 60-item BNT (r = .55), the BNT-15 (r = .54), category fluency (r = .44), and global cognitive measures (r = .35). The C-CLNT displays diagnostic utility for neurodegenerative disorders, yielding an area under the receiver operating characteristic curve (AUC) of .80 for distinguishing dementia from controls and non-dementia conditions—significantly outperforming the BNT (AUC = .64)—though its diagnostic accuracy for mild cognitive impairment (MCI) remains limited. With cross-linguistic norms and applicability spanning multiple spoken languages, the C-CLNT serves as an equitable psychometric instrument for geriatric cognitive assessment in increasingly diverse clinical settings.

Keywords

Copenhagen Cross-Linguistic Naming Test, C-CLNT, confrontation naming, anomia, cross-cultural neuropsychology, dementia diagnosis, mild cognitive impairment, MULTIMAP, psycholinguistics, geriatric assessment

Authors

The Copenhagen Cross-Linguistic Naming Test was developed by a collaborative team of clinical neuropsychologists, cognitive neurologists, and cross-cultural psychometricians:

• T. Rune Nielsen, Ph.D. – Copenhagen University Hospital–Rigshospitalet, Department of Neurology, Danish Dementia Research Centre, Copenhagen, Denmark (Corresponding Author: [email protected]).
• Bernadette Unmack Grollenberg, M.Sc. – Copenhagen University Hospital–Rigshospitalet, Department of Neurology, Danish Dementia Research Centre, Copenhagen, Denmark.
• Signe Pertou Ringkøbing, M.Sc. – Copenhagen University Hospital–Rigshospitalet, Department of Neurology, Danish Dementia Research Centre, Copenhagen, Denmark.
• Maria Özden, M.Sc. – Copenhagen University Hospital–Rigshospitalet, Department of Neurorehabilitation and Spinal Cord Injuries, The Neuroscience Centre, Copenhagen, Denmark.
• Brendan Weekes, Ph.D. – University of Hong Kong, Faculty of Education, Laboratory for Communication Science, Hong Kong SAR, China.
• Gunhild Waldemar, M.D., DMSc – Copenhagen University Hospital–Rigshospitalet, Department of Neurology, Danish Dementia Research Centre, Copenhagen, Denmark.

Purpose

The primary purpose of the Copenhagen Cross-Linguistic Naming Test (C-CLNT) is to provide an equitable, psychometrically sound assessment of visual confrontation naming in older adults from diverse ethnocultural, linguistic, and educational backgrounds presenting to memory assessment services. Confrontation naming is a core component of comprehensive neuropsychological evaluations, serving as a sensitive clinical marker for focal lesions, primary progressive aphasias, and neurodegenerative dementias—most notably Alzheimer’s disease and frontotemporal lobar degeneration.

For decades, clinical neuropsychology has relied heavily on tools developed within monolingual, Western, Educated, Industrialized, Rich, and Democratic (WEIRD) populations, predominantly the Boston Naming Test. When administered to linguistic minorities, immigrant elders, or individuals with limited formal education, legacy instruments frequently produce elevated false-positive rates for cognitive impairment. This measurement artifact arises because target items often possess disparate cultural familiarity, variable visual typicality, or divergent lexical frequencies across languages, which can lead to misdiagnoses of dementia in cognitively intact minority populations.

The C-CLNT was engineered to mitigate these diagnostic inequities. By drawing its visual stimuli from the empirically calibrated MULTIMAP repository, the test ensures that every depicted stimulus maintains high visual clarity, low cultural idiosyncrasy, and high name agreement across numerous non-cognate and cognate linguistic systems. The test was developed to operate effectively within specialized memory clinics serving multicultural populations, such as those found throughout urban European centers. It enables clinicians to differentiate genuine neurodegenerative language breakdown from life-long linguistic differences or low formal educational attainment.

Additionally, the C-CLNT incorporates both object naming (nouns) and action naming (verbs), allowing for clinical characterization of grammatical category dissociations. This distinction is diagnostically informative in separating retro-rolandic temporoparietal pathologies from frontal-striatal or motor speech degenerations. Its brevity (requiring under five minutes to administer) makes it practical for multi-tiered cognitive screening batteries, bed-side evaluations, and longitudinal tracking in diverse memory clinic settings.

Psychological Construct

The C-CLNT operationalizes the construct of visual confrontation naming—a multi-stage cognitive-linguistic process wherein a visual stimulus must be perceived, recognized, conceptually processed, lexically retrieved, and phonologically articulated. The test isolates two primary psycholinguistic subdomains within this overarching construct:

1. Concrete Object Naming (Noun Retrieval)

The object naming domain consists of 20 color illustrations depicting high-familiarity physical entities (e.g., biological specimens, tools, anatomical structures, household implements). Neurobiologically, noun retrieval relies upon the ventral visual processing stream (‘what’ pathway) traversing the occipitotemporal cortex, feeding into the anterior and ventral temporal lobes for semantic feature integration, and engaging the left middle and superior temporal gyri for lexical lemma selection. Deficits in this domain typically reflect degradation of the semantic store itself (as observed in semantic variant primary progressive aphasia) or disrupted lexical-semantic access mechanisms characteristic of early-stage Alzheimer’s disease.

2. Action Naming (Verb Retrieval)

The action naming domain consists of 10 color illustrations depicting dynamic human behaviors, movements, or transitivity interactions. Action naming imposes distinct cognitive and neural demands compared to object naming. Verbs represent relational concepts with complex argument structures (e.g., thematic roles such as agent, instrument, and patient), recruiting broader distributed neural circuits that encompass the left frontal operculum, the posterior middle temporal gyrus, and premotor cortical regions. Inclusion of action naming enables the C-CLNT to identify subtle psycholinguistic dissociations, providing greater diagnostic sensitivity to frontal and subcortical network dysfunctions seen in frontotemporal lobar degenerations, corticobasal syndrome, and progressive supranuclear palsy.

3. Cue Responsiveness and the Lexical Retrieval Cascade

The construct measured by the C-CLNT extends beyond binary item success to evaluate the locus of retrieval breakdown via a structured cueing paradigm:

• Spontaneous Retrieval: Unassisted naming within a 5-second interval reflects intact structural activation from visual input to semantic concept and rapid access to the phonological output lexicon.
• Semantic Cueing: Providing categorical or contextual semantic information tests whether perceptual degradation or high-level visual misidentification hindered lexical access. Success following a semantic prompt signifies that the patient retains the semantic concept and requires contextual activation to resolve lexical competition.
• Phonemic Cueing: Providing the initial phoneme bypasses lexical selection to directly prompt the phonological output form. Preservation of phonemic cue responsiveness suggests spared semantic knowledge accompanied by lexical access/retrieval blocking, commonly encountered in subcortical vascular cognitive impairment or normal cognitive aging, whereas failure to benefit from phonemic cues indicates frank loss of the underlying lexical-semantic representation.

Theoretical Framework

The architecture of the C-CLNT is grounded in cognitive psycholinguistic architectures of speech production, primarily drawing from Levelt’s Stage Model of Lexicalization (Levelt et al., 1999) and Dell’s Interactive Activation Theory (Dell, 1986).

1. Levelt’s Discrete-Stage Architecture

Levelt’s psycholinguistic model posits that naming an overt visual stimulus proceeds through distinct, feed-forward stages:

1. Conceptual Preparation (Macro- and Micro-planning): Visual sensory inputs undergo structural feature extraction, activating visual object representations in the inferior temporal cortex. This leads to the activation of a non-verbal lexical concept within semantic memory.
2. Lemma Selection: The activated lexical concept prompts competition among candidates in the mental lexicon. The abstract grammatical representation—the lemma—is selected, specifying syntactic category (e.g., noun vs. verb), gender, and subcategorization frames without phonological form information.
3. Phonological Encoding (Lexeme Retrieval): The selected lemma activates its corresponding morphemic and phonological word-form (the lexeme). The phonemes are sequentially organized and syllabified.
4. Articulatory Planning and Execution: The phonetic plan is translated into motor commands executed through the articulatory apparatus.

Within this framework, the C-CLNT tests the integrity of the pathway connecting conceptual preparation to lemma and lexeme retrieval. Culturally biased instruments fail because unfamiliar visual stimuli introduce static at the conceptual preparation stage, distorting lemma competition. The C-CLNT controls for visual complexity, image agreement, and concept familiarity, ensuring that any disruption reflects neurocognitive pathology rather than psycholinguistic variation.

2. Bilingual Interactive Activation and Shared Conceptual Systems

For multilingual and culturally diverse older adults, the C-CLNT operates under the framework of the Revised Hierarchical Model (Kroll & Stewart, 1994) and the Bilingual Interactive Activation Plus (BIA+) model (Dijkstra & van Heuven, 2002). These models demonstrate that multilingual individuals maintain distinct lexical stores for each language (L1, L2, Ln) connected to a common, shared semantic-conceptual memory store.

Traditional single-language naming assessments penalize bilingual individuals due to distributed lexical knowledge and cross-linguistic competition (the frequency-lag hypothesis). By selecting target concepts that demonstrate equivalent name agreement and frequency profiles across diverse linguistic systems, the C-CLNT balances psycholinguistic weight across languages. This design enables the instrument to function reliably regardless of whether testing occurs in the participant’s primary or secondary language.

Validity

The psychometric validation of the C-CLNT was conducted by Nielsen et al. (2023) in a clinical cohort of patients evaluated at the Danish Dementia Research Centre, Rigshospitalet. The validation sample encompassed healthy older control participants, individuals diagnosed with affective disorders (e.g., major depressive disorder presenting with subjective cognitive complaints), patients with mild cognitive impairment (MCI), and patients with verified dementia syndromes.

1. Construct and Convergent Validity

The construct validity of the C-CLNT was verified through comparative analyses against established neuropsychological benchmarks of language and global cognition:

• Correlation with the Boston Naming Test: The 30-item C-CLNT exhibited moderate-to-strong correlations with the standard 60-item Boston Naming Test (r = .55, p < .001) and the abbreviated 15-item BNT (r = .54, p < .001). These coefficients reflect strong construct alignment in assessing confrontation naming, while the non-redundant variance underlines the C-CLNT’s reduced sensitivity to extraneous socio-cultural bias.
• Correlation with Category Verbal Fluency: Naming performance significantly correlated with semantic category fluency (animals) at r = .44 (p < .001), indicating shared reliance on semantic memory search mechanisms.
• Correlation with Global Cognitive Functioning: Performance on the C-CLNT demonstrated statistically significant associations with broad cognitive status measures (r = .35, p < .001), confirming its expected link to general neurocognitive health.
• Equivalence of Abbreviated Forms: The 20-item object-only version showed high linear concordance with the full 30-item instrument (r = .87, p < .001), demonstrating construct preservation in the brief version.

2. Discriminant and Diagnostic Validity

Diagnostic utility was evaluated using Receiver Operating Characteristic (ROC) curve analyses to differentiate individuals with dementia from non-demented participants:

• Diagnostic Accuracy for Dementia: The 30-item C-CLNT achieved an Area Under the Curve (AUC) of .80 (95% CI [.72, .88]), demonstrating clinical utility for identifying dementia in multicultural cohorts. This performance was significantly superior (p < .001) to that of the traditional Boston Naming Test, which yielded an AUC of only .64.
• Cutoff Scores for the 30-Item Version: An optimal cutoff score of ≤ 28 points yielded a sensitivity of .75 (75%) and a specificity of .79 (79%) for diagnosing dementia.
• Diagnostic Accuracy of the 20-Item Version: The abbreviated 20-item scale demonstrated an AUC of .78 (95% CI [.70, .86]), which did not significantly differ from the diagnostic accuracy of the full 30-item battery (z = 1.21, p = .23). At an optimal cutoff score of ≤ 18 points, the 20-item version demonstrated a sensitivity of .83 (83%) and a specificity of .70 (70%).
• Limitations in Mild Cognitive Impairment (MCI): The discriminative capacity of the C-CLNT in isolating amnestic or non-amnestic MCI from cognitively intact controls was relatively limited. This finding is consistent with psycholinguistic observations that gross visual confrontation naming often remains preserved until neurodegenerative processes progress beyond focal medial temporal structures to broader neocortical language networks.

Reliability

The scale reliability of the Copenhagen Cross-Linguistic Naming Test has been evaluated across clinical and normative validation cohorts:

• Internal Consistency: The standard 30-item C-CLNT demonstrated a Cronbach’s alpha of .67. In neuropsychological confrontation naming instruments, an alpha coefficient in this range is expected. Visual naming tests balance items across differing conceptual domains, lexical frequencies, and word-lengths to cover a broad spectrum of item difficulties without redundant item content.
• Split-Half and Form Intercorrelation: The strong correlation between the 20-item object subset and the comprehensive 30-item score (r = .87) reflects high internal structural consistency across its noun and verb components.
• Inter-Rater and Scoring Stability: Because scoring is standardized around objective target responses and explicit semantic/phonemic cueing steps within fixed 5-second intervals, the C-CLNT demonstrates high inter-scorer concordance when administered by trained clinicians.

Factor Analysis

During the development and initial validation of the C-CLNT by Nielsen et al. (2023), formal exploratory factor analysis (EFA) or confirmatory factor analysis (CFA) was not reported. The items were selected through experimental psycholinguistic calibration derived from the MULTIMAP project rather than exploratory latent factor reduction.

Item selection was governed by two criteria across five European languages (Spanish, Basque, Catalan, French, English) and two non-European languages (Mandarin Chinese, Arabic):

• High Cross-Linguistic Name Agreement: Minimizing dialectal, regional, and synonym dispersion so that the visual stimulus reliably elicits a single predominant target lemma across languages.
• Harmonized Lemma Frequency: Ensuring balanced lexical frequency distributions across languages to avoid items that are common in one language but rare in another.

From a psychometric standpoint, confrontation naming tasks that sample items across a spectrum of difficulties to identify mild-to-severe anomia typically satisfy a unidimensional latent space (reflecting an overarching visual-lexical retrieval factor), or a bifactor structure separating noun and verb domains. Subsequent Item Response Theory (IRT) and structural equation modeling on larger multi-center cohorts will be valuable to formally establish parameter invariance and item characteristic curves across specific immigrant language subgroups.

Instrument / Measurement Tool

• Test Name: Copenhagen Cross-Linguistic Naming Test (C-CLNT)
• Alternative Abbreviated Version: C-CLNT 20-Item Short Form (Object Naming only)
• Test Type: Performance-based visual confrontation naming test
• Source Database: MULTIMAP (Gisbert-Muñoz et al., 2021)
• Target Population: Culturally, linguistically, and educationally diverse adults (Ages 18 to 85+ years) presenting with cognitive concerns, affective disorders, mild cognitive impairment, or suspected dementia syndromes
• Total Number of Items: 30 items (Full Scale: 20 concrete objects, 10 actions); 20 items (Abbreviated Scale: 20 concrete objects)
• Administration Duration: Less than 5 minutes (< 5 minutes)
• Languages Validated/Calibrated: Spanish, Basque, Catalan, Italian, French, English, German, Mandarin Chinese, and Arabic
• Discontinuation Rule: No discontinuation rule; all items are administered sequentially to maximize diagnostic profiling across varied semantic sub-categories
• Response and Cueing Format:
  • Stimuli are presented visually as standardized color drawings one at a time.
  • The participant is allowed 5 seconds to provide a spontaneous response.
  • If the participant fails to respond, misperceives the stimulus, or provides a superordinate category name, a standardized semantic cue is provided.
  • Participants are given 5 seconds to respond following a semantic cue or phonetic cue.
  • If the participant continues to fail after the semantic prompt, a standardized phonetic (phonemic) cue (initial phonological sound of the target word) is administered.
• Scoring Criteria:
  • The C-CLNT total score is the number of correct responses, including responses provided after semantic cues (Maximum Score = 30 for the Full Version; Maximum Score = 20 for the Abbreviated Version).
  • Responses after phonemic cues are not included in the total correct score but can be noted to provide qualitative information about naming performance.
• Clinical Cutoff Scores (Dementia Screening):
  • 30-Item Version: Total Score ≤ 28 indicates probable naming impairment consistent with dementia (Sensitivity: .75, Specificity: .79, AUC = .80).
  • 20-Item Version: Total Score ≤ 18 indicates probable naming impairment consistent with dementia (Sensitivity: .83, Specificity: .70, AUC = .78).

Permissions & Fee and Test Year

• Test Year: 2023
• Copyright and Intellectual Property: © 2023 T. Rune Nielsen, Danish Dementia Research Centre, Rigshospitalet, Copenhagen, Denmark. The underlying pictorial stimuli originate from the MULTIMAP repository (© Gisbert-Muñoz et al., 2021).
• Fee: No fee. The instrument is accessible without commercial licensing costs for non-commercial clinical, diagnostic, and academic research applications.
• Permissions and Test Acquisition: Clinicians and researchers wishing to obtain the testing stimuli, scoring sheets, and cueing manuals must contact the corresponding author:
  
  T. Rune Nielsen, Ph.D.
  Danish Dementia Research Centre, Department of Neurology
  Copenhagen University Hospital–Rigshospitalet
  Email: [email protected]

References

• Dell, G. S. (1986). A spreading-activation theory of retrieval in sentence production. Psychological Review, 93(3), 283–321. https://doi.org/10.1037/0033-295X.93.3.283
• Dijkstra, T., & van Heuven, W. J. (2002). The architecture of the bilingual word recognition system: From coding orthography to phonology. Bilingualism: Language and Cognition, 5(3), 175–197. https://doi.org/10.1017/S1366728902003012
• Gisbert-Muñoz, S., Pérez-Sánchez, M. A., & Weekes, B. S. (2021). MULTIMAP: A multilingual picture database for cognitive and psycholinguistic research. Behavior Research Methods, 53(4), 1746–1761. https://doi.org/10.3758/s13428-020-01524-7
• Kaplan, E., Goodglass, H., & Weintraub, S. (2001). Boston Naming Test (2nd ed.). Pro-Ed.
• Kroll, J. F., & Stewart, E. (1994). Category interference in translation and picture naming: Evidence for asymmetric connections between bilingual memory representations. Journal of Memory and Language, 33(2), 149–174. https://doi.org/10.1006/jmla.1994.1008
• Levelt, W. J., Roelofs, A., & Meyer, A. S. (1999). A theory of lexical access in speech production. Behavioral and Brain Sciences, 22(1), 1–38. https://doi.org/10.1017/s0140525x99001776
• Morris, J. C., Heyman, A., Mohs, R. C., Hughes, J. P., van Belle, G., Fillenbaum, G., Mellits, E. D., & Clark, C. (1988). The Consortium to Establish a Registry for Alzheimer’s Disease (CERAD). Part I. Clinical and neuropsychological assessment of Alzheimer’s disease. Neurology, 39(9), 1159–1165. https://doi.org/10.1212/wnl.39.9.1159
• Nielsen, T. R., Grollenberg, B. U., Ringkøbing, S. P., Özden, M., Weekes, B., & Waldemar, G. (2023). The Copenhagen Cross-Linguistic Naming Test (C-CLNT): Development and validation in a multicultural memory clinic population. Journal of the International Neuropsychological Society, 29(10), 911–921. https://doi.org/10.1017/S1355617723000437

Items of the Scale