PURPOSE: Spatially fractionated radiotherapy is an unconventional radiotherapy approach able to widen the therapeutic window for difficult-to-treat cases today. To unlock its full potential, accurate knowledge of the relationship between the different dosimetry and geometry parameters and the biological response is still needed. When the same beam width is used, the valley dose has been assumed to be the main parameter influencing normal tissue sparing, with peak doses having little relevance. However, a recent retrospective evaluation of preclinical data suggests peak dose plays a major role in the normal tissue sparing of minibeam radiation therapy (MBRT). The goal of this study was to experimentally validate for the first time the significance of the peak dose for normal tissue sparing in proton MBRT. MATERIALS AND METHODS: We irradiated the brains of naive rats with two different configurations of pMBRT, resulting in the same valley and average doses but different peak doses. Behavioural tests and histopathological evaluations were carried out. RESULTS: At the same valley dose, a higher peak dose (high peak-to-valley dose ratio (PVDR), larger centre-to-centre (ctc) distance) is more detrimental than a lower peak dose (low PVDR, narrower catch). In the first case, the animals exhibited some hyperactivity in locomotor and exploration activity as well as memory alterations. In addition, the highest peak dose led to a significantly higher cumulative lesion score in the histopathology evaluations than the lowest peak dose. CONCLUSIONS: While our study does not exclude the relevant role of the valley dose in tissue sparing, it does highlight the importance of peak doses, contradicting previous assumptions. Our results agree with the conclusion of a recent retrospective evaluation of preclinical studies in micro and minibeam radiation therapy.