Studying the temporal integration of visual signals is crucial to understand how time spent on different visual tasks can affect emmetropization and refractive error development. In this study we assessed the effect of interrupting positive and negative lens-imposed defocus with brief periods of unrestricted vision or darkness. A total of forty-six marmosets were treated monocularly with soft contact lenses for 4 weeks from 10 weeks of age (OD: + 5D or − 5D; OS: plano). Two control groups wore + 5D (n = 5) or − 5D (n = 13) lenses continuously for 9 h/day. Two experimental groups had lens-wear interrupted for 30 min twice/day at noon and mid-afternoon by removing lenses and monitoring vision while marmosets sat at the center of a viewing cylinder (normal vision interruption, + 5D: n = 7; − 5D: n = 8) or while they were in the dark (dark interruption, + 5D: n = 7; − 5D: n = 6). The interruption period (30 min/day) represented approx. 10% of the total stimulation time (9 h/day). On-axis refractive error (RE) and vitreous chamber depth (VCD) were measured using an autorefractor and high frequency A-scan ultrasound at baseline and after treatment. Wearing + 5D lenses continuously 9 h/day for 4 weeks induced slowed eye growth and hyperopic shifts in RE in treated relative to contralateral control eyes (relative change, VCD: − 25 ± 11 μm, p > 0.05; RE: + 1.24 ± 0.58 D, p > 0.05), whereas − 5D lens wear resulted in larger and myopic eyes (relative change, VCD: + 109 ± 24 μm, p < 0.001; RE: − 2.03 ± 0.56 D, p < 0.05), significantly different from those in the + 5D lens-treated animals (p < 0.01 for both). Interrupting lens induced defocus with periods of normal vision or darkness for approx. 10% of the treatment time affected the resulting compensation differently for myopic and hyperopic defocus. Interrupting defocus with unrestricted vision reduced − 5D defocus compensation but enhanced + 5D defocus compensation (− 5D, VCD: + 18 ± 33 μm; RE: − 0.93 ± 0.50 D, both p > 0.05; + 5D, VCD: − 86 ± 30 μm; RE: + 1.93 ± 0.50 D, both p < 0.05). Interrupting defocus with darkness also decreased − 5D defocus compensation, but had little effect on + 5D defocus compensation (− 5D, VCD: + 73 ± 34 μm, RE: − 1.13 ± 0.77 D, p > 0.05 for both; + 5D, VCD: − 10 ± 28 μm, RE: + 1.22 ± 0.50 D, p > 0.05 for both). These findings in a non-human primate model of emmetropization are similar to those described in other species and confirm a non-linear model of visual signal integration over time. This suggests a mechanism that is conserved across species and may have clinical implications for myopia management in school-aged children.