Regions of the E[e] (Quantization Error)

I don’t understand why we use the Δ/2, -Δ/2 regions for the noise error of a uniform quantizer. For being uniform quantizer, all quantization regions are of equal size (Δ)

enter image description here