The correspondence between the answers to the Q16 questions regarding memory and attention-concentration and relevant neurobehavioral performance test scores has been evaluated. The sensitivity, specificity and diagnostic validity of Q16 have been assessed, taking the relevant neurobehavioral test score as a reference diagnostic criterion, the lower quartile of performance being considered as a poor response. The group under study consisted of 74 volunteers (24 females), aged 40 years on average (SD:7.5) and recruited among styrene-exposed workers and healthy controls. The test battery included the logical memory (short- and long-term) and the verbal learning (short- and long-term) tests of the Wechsler Adult Intelligence Scale (WAIS). The answers to the Q16 questions were poorly related to the performance: self-perceived forgetfulness showed a limited agreement with the long-term logical memory test (r=-0.23, p<0.05). The number of false negatives (no symptom but low test scores) was generally high, giving rise to a very low sensitivity of the questionnaire, despite a relatively high specificity. Accordingly, the positive diagnostic validity was low (<30%), whereas the negative diagnostic validity was high (>80%). Different methods used to investigate subtle neurological changes give rise to inconsistencies between self-perceived disturbances and objective measurements of relevant functions. Owing to its low sensitivity and positive diagnostic value, the Q16 cannot be recommended as a screening tool among workers occupationally exposed to neurotoxic chemicals.