Vasopressin injections impair working memory in a delayed matching to sample task in rats

Effects of vasopressin were measured in a nonspatial working memory task: food-reinforced, delayed matching to sample. Subcutaneous injections of 0.2 microgram of lysine vasopressin (LVP) or saline were alternately administered to Sprague-Dawley rats after the presentation of the sample and compared to the effect of the same treatments given to a yoked control group of rats. Different durations of sample presentation (5 and 30 min) and various retention intervals (10 min, 3 h, 24 h) were selected. The results showed that LVP never facilitated retention performance: there was no improvement under conditions of weak memory (short presentation of the sample and long retention interval); moreover LVP abolished the facilitation normally obtained when either the length of the sample presentation was longer or the duration of the retention interval was shorter. The performance after the injection of the peptide was differentially impaired, according to brightness of the sample which had been presented: after the presentation of the white box, LVP injections lead to more errors and after that of the black one the treatment induced an increase in latencies of response. All these data may suggest that the physiological consequences of hormonal modifications triggered by the LVP injection are processed along with the stimuli of the learning episode and interfere with the learned positive value of the sample.