Microelectrode biosensors are a promising technique to probe the brain interstitial fluid and estimate the extracellular concentration of neurotransmitters like glutamate. Their selectivity is largely based on maintaining high substrate specificity for the enzymes immobilized on microelectrodes. However, the effect of enzyme immobilization on substrate specificity is poorly understood. Furthermore, the accuracy of biosensor measurements for brain biological extracts has not been reliably established in comparison with conventional analytical techniques. In this study, microelectrode biosensors were prepared using different enzyme immobilization methods, including glutaraldehyde, a conventional cross-linker, and poly(ethylene glycol) diglycidyl ether (PEGDE), a milder immobilization reagent. Glutaraldehyde, but not PEGDE, significantly decreased the apparent substrate specificity of glutamate and glucose oxidase. For glutaraldehyde prepared biosensors, detection of secondary substrates by glutamate oxidase increased, resulting in a significant overestimate of glutamate levels. This effect was not observed with PEGDE-based biosensors, and when brain microdialysates were analyzed, the levels of glutamate detected by biosensors were consistent with those detected by capillary electrophoresis. In addition, basal concentrations of glutamate detected in vivo were approximately 10-fold lower than the levels detected with glutaraldehyde-based biosensors (e.g., 1.2 μM vs 16 μM, respectively). Overall, enzyme immobilization can significantly impact substrate specificity, and PEGDE is well-suited for the preparation of stable and selective biosensors. This development questions some of the previous biosensor studies aimed at detecting glutamate in the brain and opens new possibilities for specific neurotransmitter detection.