Remote photoplethysmography (rPPG) has traditionally relied on facial videos for contactless cardiovascular monitoring; however, the anterior neck presents a promising alternative due to the superficial proximity of the carotid arteries. Despite this advantage, rPPG signals in the neck region are often affected by spatial variability in signal quality. This paper proposes a systematic framework to characterize the spatial distribution of rPPG signals across the neck. By partitioning the neck region of interest (ROI) into a 10 × 20 grid, we independently extract rPPG signals from 200 discrete spatial cells. A composite quality metric, S, is developed by combining temporally normalized signal energy with spectral signal-to-noise ratio (SNR). Results reveal pronounced spatial heterogeneity and subject-specific variability in signal quality. Bland-Altman analysis demonstrates that combined rPPG signal from the cells in the upper-quartile composite quality (S>Q3) regions achieve high agreement with gold-standard ECG (mean bias: 1.09 bpm; limits of agreement: -6.83 to 9.01 bpm), whereas those within the lower-quartile S (S<Q1) regions exhibit a dramatic degradation in accuracy (limits of agreement: -26.09 to 28.62 bpm). These findings underline the necessity of spatially adaptive signal selection and provide motivations for subject-specific ROI optimization in neck-based rPPG.