Recent evidence has indicated that NSAIDs might have direct effects on CNS tissue in addition to their classical inhibitory action on COX enzymes. This thesis addresses this hypothesis using electrophysiological and behavioural techniques. The effects of fenamate and other NSAIDs on native neuronal GABA(_A), 5-HT(_3), nicotinic ACh, P2x and strychinine-sensitive glycine receptors, expressed on isolated vagus or optic nerves, was investigated using an extra-cellular recording technique. The fenamate NSAID, mefenamic acid (MFA) potentiated GABA (10µM)- evoked responses in the vagus nerve. Application of MFA also resulted in non-competitive inhibition of 5-HT-and a,βMeATP- evoked responses. Non-competitive like inhibition was also observed with flufenamic acid on DMPP- and a,βMeATP- evoked responses and with meclofenamic acid on GABA- evoked responses. Non-fenamate NSAIDs, including aspirin, did not significantly modulate the GABA(_A), 5-HT(_3), nicotinic ACh, P2x or glycine receptors. The cognitive and behavioural effects of fenamates and other NASIDs were then investigated. MFA (5-20mg/kg) caused a significant dose- and time-dependent enhancement in the non-spatial object discrimination working memory task when compared to saline controls. The enhancement observed with MFA was greater than that of the cognitive enhancer piracetam. This enhancement was not due to a change in non-mnemonic processes such as arousal, anxiety or locomotion. MFA also enhanced rats' performance in the spatial object location working memory task. The fenamate NSAID, meclofenamic acid (20mg/kg) mimicked the effect of MFA, but the non-fenamate NSAIDs aspirin and ibuprofen, did not enhance object discrimination indicating that these cognitive effects are not via inhibition of COX. The GABA(_A) receptor modulators diazepam, bicuculline and loreclezole, did not replicate the effect of MFA on object discrimination, suggesting that its effects do not depend entirely on the GABAa receptor. Scopolamine (0.25-lmg/kg) significantly impaired object discrimination in a dose-dependent manner. This action could be fully reversed by co-treatment with MFA (20mg/kg).In the T-maze task, MFA (20mg/kg) decreased the number of arm entry errors and days taken to reach criterion. The number of arm entry errors made when a 5-minute intra-trial interval was introduced was also significantly reduced by MFA compared with saline treated animals. In the radial maze, MFA (20mg/kg) did not decrease the number of never baited arm entries to reach criterion. However MFA did significantly reduce the number of re-entry errors to baited arms, compared to controls, when an intra-trial delay (10-30 sees) was introduced. These results support the hypothesis that MFA enhances spatial working memory and that these effects are not task-specific. Overall, the data in this thesis show that fenamate NSAIDs can directly modulate native neuronal ligand-gated ion channels and that MFA can enhance working memory in normal and scopolamine-impaired rats. These results suggest additional pharmacological potential for certain fenamate NSAIDs.