An abnormality in glutamate function has been implicated in the neural substrate of depressive disorders. To investigate this in rats, the Porsolt swim test was used to assess the role of glutamate in the nucleus accumbens. Glutamate injected into the nucleus accumbens dose-dependently decreased swimming time on the test day (day 2), whereas N-methyl-D-aspartate antagonists dizocilpine and 2-amino-5-phosphonovalerate increased swimming, like an antidepressant. Dizocilpine injected before the conditioning trial (day 1) did not modify the swimming times during the first day but abolished behavioral depression on day 2. Microdialysis coupled to capillary-zone electrophoresis was then used to determine in vivo changes in glutamate release in 1-min samples during the swim test. On day 1, glutamate increased significantly and reached a maximum of 222% after 3 min of swimming. On day 2, baseline glutamate levels were back to normal, but when the animal was placed in the water, glutamate increased to 419% during the first minute, and the animals swam significantly less. For comparison, tail pinch on consecutive days was used as a nonspecific, repeated stressor while accumbens glutamate levels were measured. Tail pinch on the first day increased glutamate similar to the effect obtained during the first day of swimming; however, a second day of tail pinch decreased glutamate levels, instead of the potentiated response observed during the second day of swimming. These results show that accumbens glutamate plays a role in causing the behavioral aspects of depressed behavior as modeled in the swim test. The accumbens may be a potential site of action for drugs that alter behavioral depression. (C) 2003 IBRO. Published by Elsevier Science Ltd. All rights reserved.