With m-Dinitrobenzene (m-DNB) as a selected model neurotoxicant, we demonstrate how to assess neurotoxicity, using morphology based measurement of neurite degeneration, in a conventional “full-contact” and a modern “restricted-contact” co-culture of rat cortical neurons and astrocytes. In the “full-contact” co-culture, neurons and astrocytes in complete physical contact are “globally” exposed to m-DNB. A newly emergent “restricted-contact” co-culture is attained with a microfluidic device that polarizes neuron somas and neurites into separate compartments, and the neurite compartment is “selectively” exposed to m-DNB. Morphometric analysis of the neuronal area revealed that m-DNB exposure produced no significant change in mean neuronal cell area in “full-contact” co-cultures, whereas a significant decrease was observed for neuron monocultures. Neurite elaboration into a neurite exclusive compartment in a compartmentalized microfluidic device, for both monocultures (no astrocytes) and “restricted” co-cultures (astrocytes touching neurites), decreased with exposure to increasing concentrations of m-DNB, but the average neurite area was higher in co-cultures. By using co-culture systems that more closely approach biological and architectural complexities, and the directionality of exposure found in the brain, this study provides a methodological foundation for unraveling the role of physical contact between astrocytes and neurons in mitigating the toxic effects of chemicals such as m-DNB.