A large-scale field survey was undertaken in the Neste system (South-West France) to investigate the impact of pesticide inputs on the structure of riverine benthic diatom communities. A total of 18 sites in the "Coteaux de Gascogne" study area were sampled during periods of pesticide use over two consecutive years (2005-2006). The typology of the diatoms collected was determined using artificial neural networks generating patterns in diatom community composition that indicated the species influenced by pesticide inputs, combined with organic pollution. Small, pioneer species of the Achnanthales group were more frequent in the pesticide-contaminated assemblages, whereas slower colonizers were more often found in the sites where disturbance was lower. The Phytopixal approach, a means to assess the spatial contamination potential, provided valuable information about pesticide exposure and was a successful means of discriminating diatom assemblages, suggesting that a global estimate of pesticide pollution risk, although rough, would be more appropriate and more representative than sporadic pesticide analyses.