Lorazepam but not diazepam impairs identification of pictures on the basis of specific contour fragments.

We tested the effects of lorazepam 0.038 mg/kg and diazepam 0.3 mg/kg on the identification of pictures of everyday objects that were specifically modified to examine the role of different parts of the external contour. By pressing the space bar of a computer keyboard, observers could add 1% of the total contour of each picture until it was recognized. Identification thresholds were measured in three display conditions, depending on where along the contour the addition of contour pixels started. In the minima condition, stimuli were initially displayed with only minima (i.e., locations along the contour where negative curvature is strongest); all parts with negative curvature were then built up gradually from the minima and only later on were the fragments with positive curvature shown until the contour became closed at the maxima (i.e., locations where positive curvature is strongest). In the maxima condition, initially only the maxima were displayed, with all positive contour built up first and then the negative curvature until the minima were reached. In the inflections condition, the points along the contour shown first were inflections (i.e., points where curvature is locally zero because the sign of curvature changes there) and contour was built up by adding parts of positive and negative curvature at both sides of each inflection until the extrema (minima and maxima) were reached to close the contour of the picture. In general, picture identification was more difficult (i.e., a larger portion of the contour was required) in the minima condition than in the maxima and the inflections conditions. The diazepam group did not differ significantly from the placebo group, while the lorazepam group had significantly lower performance in all three display conditions. Results are discussed in relation to previous research showing impaired perceptual integration and impaired implicit memory under lorazepam influence.