Socio-economic constructs and urban topology are crucial drivers of human mobility patterns. During the COVID-19 pandemic, these patterns were re-shaped in their main two components: the spatial dimension represented by the daily travelled distance, and the temporal dimension expressed as the synchronisation time of commuting routines. Leveraging location-based data from de-identified mobile phone users, we observed that during lockdowns restrictions, the decrease of spatial mobility is interwoven with the emergence of asynchronous mobility dynamics. The lifting of restriction in urban mobility allowed a faster recovery of the spatial dimension compared to the temporal one. Moreover, the recovery in mobility was different depending on urbanisation levels and economic stratification. In rural and low-income areas, the spatial mobility dimension suffered a more significant disruption when compared to urbanised and high-income areas. In contrast, the temporal dimension was more affected in urbanised and high-income areas than in rural and low-income areas.