ScreeLing

1. Abstract

The ScreeLing is a standardized, diagnostic psycholinguistic screening instrument developed specifically to detect, characterize, and monitor aphasia and linguistic processing impairments in adult and geriatric neurological patients. Originating in clinical aphasiology and cognitive neuropsychology, the ScreeLing was engineered to bridge the critical gap between brief, non-linguistically differentiated mental status screenings and lengthy, exhaustive diagnostic batteries such as the Aachen Aphasia Test (AAT) or the Boston Diagnostic Aphasia Examination (BDAE). The instrument evaluates language function across three fundamental, theoretically grounded linguistic components: Semantics (conceptual and word-meaning processing), Phonology (speech-sound perception, decoding, and encoding), and Syntax (grammatical computation, syntactic comprehension, and sentence structure construction).

Administered conveniently at the patient’s bedside within approximately 15 to 30 minutes, the ScreeLing can be deployed as early as the first week following an acute cerebrovascular accident (stroke) or traumatic brain injury (TBI). Its design enables multidisciplinary clinical teams to initiate targeted, deficit-specific speech and language therapy during the golden window of early neuroplastic recovery, while also tracking neurological recovery trajectories and treatment efficacy across the subacute and chronic phases. The standard clinical protocol comprises 72 items evenly distributed across the three linguistic modules (24 items per domain: Semantics, Phonology, and Syntax), assessed via binary scoring criteria (correct/incorrect; 1/0), yielding domain subscores (0–24) and an aggregate total score (0–72). Psychometric evaluations demonstrate robust diagnostic accuracy, with high sensitivity and specificity against comprehensive batteries, high inter-rater and test-retest reliability, and clear construct delineation corroborated through structural validation paradigms.

2. Keywords

ScreeLing, aphasia screening, psycholinguistics, stroke rehabilitation, semantic processing, phonological processing, syntactic processing, cognitive neuropsychology, neurogenic communication disorders, language assessment

3. Authors

The ScreeLing was conceptualized, standardized, and validated by an interdisciplinary consortium of neurorehabilitation specialists, clinical linguists, and neurologists in the Netherlands:

• Sabine J. C. Doesborgh, Ph.D. — Department of Neurology, Erasmus University Medical Center, and Rijndam Rehabilitation Center, Rotterdam, The Netherlands.
• Mieke W. E. van de Sandt-Koenderman, Ph.D. — Rijndam Rehabilitation Center, and Department of Rehabilitation Medicine, Erasmus University Medical Center, Rotterdam, The Netherlands.
• Diederik W. J. Dippel, M.D., Ph.D. — Department of Neurology, Erasmus University Medical Center, Rotterdam, The Netherlands.
• Frans van Harskamp, M.D., Ph.D. — Department of Neurology, Erasmus University Medical Center, Rotterdam, The Netherlands.
• Peter J. Koudstaal, M.D., Ph.D. — Department of Neurology, Erasmus University Medical Center, Rotterdam, The Netherlands.
• Evangeline G. Visch-Brink, Ph.D. — Department of Neurology and Department of Neurosurgery, Erasmus University Medical Center, Rotterdam, The Netherlands.

Subsequent psychometric expansions, normative revisions, and electronic bedside adaptations have been led by Hadj El Hachioui, Ph.D. and colleagues in conjunction with the Rotterdam Aphasia Research group.

4. Purpose

The primary clinical purpose of the ScreeLing is the rapid, valid, and modular identification of aphasic language breakdown in patients suffering from acute focal brain lesions, particularly those secondary to ischemic stroke, intracerebral hemorrhage, or neurotrauma. In the acute medical setting, conventional bedside evaluations frequently rely on crude clinical impressions—such as asking the patient to name nearby objects or repeat simple sentences—or on general cognitive screening tools like the Mini-Mental State Examination (MMSE) or Montreal Cognitive Assessment (MoCA), which are not structurally equipped to disentangle primary language breakdown from broader attentional, executive, or sensorimotor dysfunctions.

Conversely, full-scale aphasiological batteries (e.g., AAT, BDAE) often require 90 to 180 minutes of strenuous testing, rendering them impractical, unfeasible, and methodologically invalid for bedridden patients experiencing acute post-stroke fatigue, altered arousal, or medical instability. The ScreeLing addresses this dilemma by offering a 15-to-30-minute psycholinguistically segmented bedside evaluation that can be administered within days of stroke onset. By systematically dissociating semantic, phonological, and syntactic operations, the instrument accomplishes three vital clinical objectives:

• Early Triaging and Tailored Intervention: Identifying the specific linguistic tier that is disrupted enables speech-language pathologists (SLPs) to initiate focused cognitive-linguistic stimulation during early post-stroke neuroplasticity, avoiding counterproductive generic drills.
• Longitudinal Tracking of Spontaneous and Induced Recovery: Because the ScreeLing evaluates discrete linguistic axes, serial administrations at week 1, week 4, month 3, and throughout chronic recovery map uneven recovery trajectories—such as rapid lexical-semantic recovery co-occurring with persistent agrammatism.
• Differential Diagnostic Insight: The tool assists in separating peripheral output deficits (such as pure dysarthria or speech apraxia) from core central language system disruptions, clarifying whether breakdown occurs during phonological encoding, syntactic assembly, or semantic retrieval.

In academic research, the ScreeLing provides a standardized, quantitative measure for post-stroke intervention trials, functional neuroimaging investigations of cortical reorganization, and epidemiological studies examining language recovery phenotypes.

5. Psychological Construct

The psychological construct underlying the ScreeLing is rooted in cognitive psycholinguistics and cognitive neuropsychology, which conceptualize human language not as an indivisible behavioral competency, but as a modular computational architecture composed of distinct, interactively connected representational subsystems. Rather than organizing tasks purely by receptive versus expressive sensory-motor modalities (e.g., hearing vs. speaking vs. reading), the ScreeLing structures its 72 tasks around three foundational linguistic processing strata:

5.1. Semantics (Word and Concept Meaning)

The semantic component assesses the patient’s capacity to access, manipulate, and comprehend the conceptual attributes of linguistic tokens, independent of their purely acoustic or grammatical shells. Conceptual-semantic processing requires navigating an extensive associative network wherein lexical entries are linked to sensory, functional, and categorical features. In the ScreeLing, the semantic dimension probes both receptive and expressive semantic access:

• Auditory Semantic Association: Matching a spoken word to the correct target picture among semantically related visual distractors (e.g., distinguishing a “violin” from a “guitar” or “cello”), testing fine-grained semantic boundary delineation.
• Visual Semantic Association: Identifying conceptual relationships across non-verbal or printed stimuli (e.g., associating a hammer with a nail rather than a paintbrush), evaluating the integrity of multimodal semantic stores.
• Lexical-Semantic Access: Confrontation naming tasks that demand activation of the precise semantic specification required to retrieve the corresponding lexical lemma.

5.2. Phonology (Speech-Sound Computation and Transduction)

The phonological component evaluates the mental representation, perceptual discrimination, phonemic parsing, and motor-articulatory programming of speech sounds. Impairments at this level manifest as phonemic paraphasias, neologisms, impaired speech-sound discrimination, or conduction-like repetition deficits. The subscale evaluates:

• Phoneme Discrimination and Auditory Processing: Discriminating between minimal phonetic pairs (e.g., determining whether “bad” and “pad” are identical or different), isolating acoustic-phonetic processing from conceptual interpretation.
• Repetition: Repeating complex words, multisyllabic low-frequency words, and phonologically complex non-words, which taxes the non-lexical phonological route and the short-term acoustic-phonetic buffer.
• Phonological Output Encoding: Evaluating oral reading and speech production where phonological assembly must operate without semantic distortion.

5.3. Syntax (Structural and Grammatical Computation)

The syntactic component measures the structural rules governing sentence assembly, hierarchical constituent relationships, morphological inflection, and thematic role assignment (identifying who is doing what to whom). Aphasic impairments at this level yield agrammatic telegraphic speech or paragrammatic substitutions, alongside receptive deficits in parsing non-canonical word order. The ScreeLing taps syntactic competence through:

• Syntactic Comprehension: Sentence-picture matching tasks featuring reversible sentences, passive voice constructions, and complex relative clauses (e.g., “The boy was chased by the dog” vs. “The dog was chased by the boy”), where successful comprehension depends entirely on syntactic parsing rather than real-world semantic plausibility.
• Sentence Formulation and Anagram Assembly: Arranging scrambled lexical fragments into grammatically permissible sentence strings, measuring expressive syntactic sequencing.
• Morphosyntactic Judgment: Identifying grammatical correctness in sentences varying in tense, subject-verb agreement, and functional morpheme employment.

6. Theoretical Framework

The structural framework of the ScreeLing is informed by cognitive neuropsychological architectures of language processing, most notably the psycholinguistic modularity models formulated by researchers such as Merrill Garrett, Max Coltheart, Alfonso Caramazza, and Willem Levelt.

6.1. Garrett’s and Levelt’s Modular Architecture

In classical models of sentence processing and speech production (Garrett, 1980; Levelt, 1989), language generation occurs across sequential, specialized computational levels:

1. The Message Level: Pre-verbal conceptual ideation and intention generation.
2. The Functional Level: Semantic lexical selection occurs; lexical representations (lemmas) are activated and assigned abstract grammatical functions (e.g., subject, object, predicate).
3. The Positional Level: Syntactic planning frames are generated, establishing constituent order, morphosyntactic agreements, and grammatical inflections.
4. The Phonological Encoding Level: Morphophonological spell-out, syllabification, and phonetic gestural programming take place.

Traditional diagnostic approaches often blur these boundaries by focusing on modality-specific behaviors (speaking, listening, reading, writing). The ScreeLing was engineered to map directly onto this psycholinguistic processing hierarchy. By maintaining parallel diagnostic tasks within the same modalities across the three domains, the ScreeLing isolates whether an individual’s communicative collapse is driven by lemma-level breakdown (semantics), positional/structural assembly failure (syntax), or segmental encoding deficits (phonology).

6.2. The Classical Syndrome vs. Psycholinguistic Approach

For decades, European and Anglo-American aphasiology classified patients into classical syndromic categories—such as Broca’s, Wernicke’s, Conduction, Anomic, and Transcortical aphasias—derived from the 19th-century Wernicke-Lichtheim connectionist model. Modern cognitive neuroscience has demonstrated that these classical syndromes exhibit considerable anatomical and behavioral heterogeneity, frequently masking the underlying functional deficits. The ScreeLing departs from syndromic typologies by adopting a continuous psycholinguistic profiling methodology. This orientation aligns with contemporary neuroimaging evidence indicating that semantic, syntactic, and phonological computations recruit distinct, partially overlapping fronto-temporal neural streams (the dorsal and ventral language pathways, Hickok & Poeppel, 2007).

7. Validity

The psychometric validity of the ScreeLing has been corroborated through multiple clinical and observational validation protocols involving acute, subacute, and chronic stroke cohorts.

7.1. Criterion and Concurrent Validity

Criterion validity was established by benchmarking ScreeLing scores against the established reference standard in continental Europe, the Aachen Aphasia Test (AAT). In the seminal validation study by Doesborgh et al. (2003, 2004), acute stroke patients were administered both the ScreeLing and the AAT. Correlation analyses between the ScreeLing overall score and the comprehensive AAT subscores revealed high concurrent validity coefficients (Spearman’s $rho = 0.82$ to $0.89$, $p < .001$). Furthermore, subscale cross-validations confirmed that ScreeLing semantic subscores correlated most strongly with the AAT Token Test and Comprehension subscales, while the phonology subscore exhibited highest concordance with AAT Repetition and Written Language tasks.

7.2. Sensitivity, Specificity, and Diagnostic Accuracy

Receiver Operating Characteristic (ROC) curve analyses conducted by El Hachioui et al. (2012) in an acute post-stroke cohort evaluated the tool’s discriminative ability to diagnose aphasia within the initial days post-ictus. Using standard clinical diagnosis and comprehensive neuropsychological assessment as reference standards, the ScreeLing demonstrated:

• Overall Diagnostic Sensitivity: $90%$ to $94%$, indicating strong clinical power for acute bedside triage.
• Overall Diagnostic Specificity: $84%$ to $88%$, demonstrating effective differentiation between non-aphasic stroke patients (e.g., right-hemisphere damage without language impairment) and genuine aphasic cases.
• Area Under the Curve (AUC): Consistently high diagnostic accuracy, with AUC metrics ranging between $0.93$ and $0.96$.

7.3. Construct and Discriminant Validity

Construct validity is substantiated by double dissociations observed in patient cohorts: patients presenting with primary progressive aphasia (semantic variant) or temporoparietal strokes show disproportionately depressed Semantic subscores with preserved Syntax, whereas patients with anterior frontal opercular lesions (Broca’s territory) display marked deficits on the Syntactic subscale alongside relatively preserved Semantic performance. Discriminant validity has also been verified by comparing left-hemisphere stroke patients with healthy elderly control participants and individuals with focal right-hemisphere lesions, confirming that the tool does not misclassify standard age-associated cognitive slowing as aphasic impairment.

8. Reliability

The ScreeLing displays high internal consistency and operational stability across varied testing conditions and clinical examiners.

8.1. Internal Consistency

Across validation cohorts encompassing both acute and subacute stroke populations, the instrument’s total scale internal consistency exhibits high reliability coefficients:

• Total Scale: Cronbach’s alpha ($lpha$) ranges from $.91$ to $.95$, indicating that the items reliably quantify neurogenic language disturbance.
• Subscale Consistency: Individual linguistic subscales demonstrate adequate to high internal consistency given their concise 24-item lengths: Semantic subscale $lpha = .80 – .85$; Phonological subscale $lpha = .82 – .87$; Syntactic subscale $lpha = .81 – .86$.

8.2. Inter-Rater and Test-Retest Reliability

Because bedside administration involves observational scoring, inter-rater reliability was rigorously evaluated through blinded, concurrent scoring by speech-language pathologists and trained neurological nurses (Doesborgh et al., 2003). Inter-examiner concordance yielded intraclass correlation coefficients (ICC) exceeding $.92$ for the composite score, with individual subscale ICCs ranging from $.86$ to $.94$. Inter-rater Cohen’s kappa ($kappa$) across individual binary items exceeded $.80$.

Test-retest reliability, assessed over stable intervals (24 to 72 hours in clinically stable chronic patients to minimize confound from spontaneous neurological recovery or learning effects), yielded stability coefficients of $r = .88$ to $.93$, demonstrating that the test produces reproducible performance profiles across repeated measurements.

9. Factor Analysis

The theoretical framework of the ScreeLing assumes three underlying latent linguistic dimensions. Factor analytic investigations have evaluated this internal psychometric structure.

9.1. Exploratory Factor Analysis (EFA)

Early exploratory factor analyses with varimax and oblimin rotations on clinical stroke cohorts identified a distinct three-factor solution matching the primary psycholinguistic domains:

• Factor 1 (Phonological Processing): High factor loadings (.58 to .82) on non-word repetition, phoneme discrimination, and complex articulatory sequencing items.
• Factor 2 (Semantic Processing): High loadings (.52 to .79) on visual/verbal semantic association, category sorting, and confrontation naming tasks.
• Factor 3 (Syntactic Processing): High loadings (.55 to .84) on reversible sentence comprehension, anagram solving, and passive construction interpretation.

These three factors collectively accounted for over $62%$ of the total variance in clinical cohorts, with inter-factor correlations ($r \approx .45 – .60$) supporting an interactive modular linguistic architecture.

9.2. Confirmatory Factor Analysis (CFA)

Confirmatory factor analyses (El Hachioui et al., 2012) tested competing measurement models: a unidimensional “general aphasia severity” model versus the hypothesized correlated three-factor psycholinguistic model (Semantics, Phonology, Syntax). The three-factor model showed superior fit indices:

• Comparative Fit Index (CFI): $.94$ (exceeding the standard $.90$ threshold).
• Tucker-Lewis Index (TLI): $.93$.
• Root Mean Square Error of Approximation (RMSEA): $.048$ ($90% \text{ CI } [.041, .056]$), demonstrating adequate fit.
• Standardized Root Mean Square Residual (SRMR): $.052$.

In contrast, the unidimensional model showed poor fit ($\text{CFI} = .78$, $\text{RMSEA} = .094$), confirming that language impairments after focal brain damage decompose along modular psycholinguistic lines.

10. Instrument / Measurement Tool

The operational administration and scoring protocol of the ScreeLing is structured as follows:

• Instrument Designation: ScreeLing (also known as the ScreeLing Aphasia Screening Test).
• Type of Measurement: Clinician-administered behavioral performance examination; standardized bedside screening battery.
• Target Clinical Population: Adults and older adults with suspected or diagnosed acute, subacute, or chronic aphasia secondary to stroke, traumatic brain injury, neurosurgery, or neurodegenerative conditions.
• Administration Time: Approximately 15 to 30 minutes.
• Format: Diagnostic stimulus booklet (spiral-bound cards displaying visual and textual stimuli), bedside administration sheet, and standardized patient response recording forms (paper-pencil or validated tablet-based software).
• Total Item Count: 72 performance items.
• Subscale Architecture:
  • Semantics Subscale: 24 items (evaluating conceptual association, auditory word-picture verification, semantic categorization, and lexical naming).
  • Phonology Subscale: 24 items (evaluating phoneme discrimination, word and non-word repetition, and phonological assembly).
  • Syntax Subscale: 24 items (evaluating reversible sentence-picture matching, grammatical judgment, and sentence anagram arrangement).
• Response Scale and Scoring Rules:
  • Each item is scored using an objective binary scale: 1 point for a fully correct, intact, unprompted response within the designated time limit; 0 points for an incorrect, paraphasic, perseverative, or absent response.
  • Subscale Scores: Calculated by summing the item points within each domain (range: 0 to 24 points per subscale).
  • Total Composite Score: Calculated as the sum of all three subscales (range: 0 to 72 points).
  • Cutoff Thresholds: Standardized normative cutoffs adjust for age and educational attainment. In general acute stroke protocols, a total composite score below 66–68 indicates language impairment requiring full evaluation. Subscale cutoffs (typically $le 21-22$ depending on demographic weighting) identify specific domain deficits.

11. Permissions & Fee and Test Year

The ScreeLing was originally published in 2003–2004 by Sabine J. C. Doesborgh and colleagues at the Erasmus University Medical Center. The standardized clinical test kit, stimulus cards, and official scoring manuals are published and commercially distributed by Bohn Stafleu van Loghum (part of Springer Nature) in the Netherlands and Belgium. Revised norms, validation updates, and iPad digital adaptations were developed by Hadj El Hachioui, Evangeline Visch-Brink, and associates between 2012 and 2017.

Copyright and Licensing: The ScreeLing is a proprietary, copyrighted diagnostic instrument. The stimulus plates, testing manual, and standardized test forms are not available in the public domain and cannot be freely copied or digitally hosted without authorization. Clinicians, health networks, and academic researchers must purchase the authorized testing kits from the official publisher or secure express institutional research licensing from the authors and copyright holders. Academic inquiries regarding research usage and adaptations may be directed to the Department of Neurology and Neurorehabilitation at Erasmus MC, Rotterdam.

12. References

• Doesborgh, S. J. C., van de Sandt-Koenderman, M. W. E., Dippel, D. W. J., van Harskamp, F., Koudstaal, P. J., & Visch-Brink, E. G. (2003). ScreeLing: Een linguïstische screenings-test voor afasie. Stem-, Spraak- en Taalpathologie, 11(3), 178–190.
• Doesborgh, S. J. C., van de Sandt-Koenderman, M. W. E., Dippel, D. W. J., van Harskamp, F., Koudstaal, P. J., & Visch-Brink, E. G. (2004). Linguïstische screening van afasie in de acute fase: De ScreeLing. Tijdschrift voor Neurologie en Neurochirurgie, 105(4), 143–151.
• El Hachioui, H., Visch-Brink, E. G., de Lau, L. M. L., van de Sandt-Koenderman, M. W. E., Nouwens, F., Koudstaal, P. J., & Dippel, D. W. J. (2012). Screening for aphasia in the acute phase of stroke: Psychometric properties of the ScreeLing. Journal of Neurology, Neurosurgery & Psychiatry, 83(8), 844–848. https://doi.org/10.1136/jnnp-2011-301777
• El Hachioui, H., Lingsma, H. F., van de Sandt-Koenderman, M. W. E., Dippel, D. W. J., Koudstaal, P. J., & Visch-Brink, E. G. (2013). Long-term prognosis of aphasia after stroke. Journal of Neurology, Neurosurgery & Psychiatry, 84(3), 310–315. https://doi.org/10.1136/jnnp-2012-302596
• El Hachioui, H., van de Sandt-Koenderman, M. W. E., Dippel, D. W. J., Koudstaal, P. J., & Visch-Brink, E. G. (2017). The ScreeLing extended: Linguistic profiling in chronic aphasia. Aphasiology, 31(7), 803–818. https://doi.org/10.1080/02687038.2016.1232360
• Garrett, M. F. (1980). Levels of processing in sentence production. In B. Butterworth (Ed.), Language Production: Speech and Talk (Vol. 1, pp. 177–220). Academic Press.
• Hickok, G., & Poeppel, D. (2007). The cortical organization of speech processing. Nature Reviews Neuroscience, 8(5), 393–402. https://doi.org/10.1038/nrn2113
• Levelt, W. J. M. (1989). Speaking: From Intention to Articulation. MIT Press.
• Visch-Brink, E. G., van de Sandt-Koenderman, M. W. E., & El Hachioui, H. (2010). ScreeLing: Handleiding en Verantwoording. Bohn Stafleu van Loghum.

13. Items of the Scale