Linguistic deficits in the acute phase of stroke

Background and Purpose For the diagnosis of aphasia early after stroke, several screening tests are available to support clinical judgment. None of these tests enables the clinician to assess the underlying linguistic deficits, i.e. semantic, phonological and syntactic deficits, which provides indispensable information for early therapeutic decisions. The ScreeLing was designed as a screening test to detect semantic, phonological and syntactic deficits. The ScreeLing's sensitivity, specificity and accuracy in detecting aphasia and semantic, phonological and syntactic deficits were determined. Methods The ScreeLing was validated in an acute stroke population against a combined reference diagnosis of aphasia (aphasia according to at least two of the following measures: neurologist's judgment, linguist's judgment, Tokentest-score). The three ScreeLing subtests were validated in the aphasic population against the presence or absence of a semantic, phonological and/or syntactic deficit according to an experienced clinical linguist. Results From a consecutive series of 215 stroke patients, 63 patients were included. The ScreeLing was an accurate test for the detection of aphasia (0.92), with a sensitivity of 86 % and specificity of 96 %. Sensitivity of subtests was 62 % for semantics, 54 % for phonology and 42 % for syntax. Specificity was 100 % for semantics and phonology and 80 % for syntax, and accuracy 0.84 for semantics, 0.87 for phonology and 0.64 for syntax. Conclusions The ScreeLing is an accurate test that can be easily administered and scored to detect aphasia in the first weeks after stroke. Furthermore, the ScreeLing is suitable for revealing underlying linguistic deficits, especially semantic and phonological deficits.