Deep Learning Sleep (DLS)

IXICO's DLS (Deep Learning Sleep) algorithm uses a convolutional neural network (CNN) to predict sleep from wrist-worn actigraphy. Compared to traditional, minute-based scoring of activity, this novel AI approach allows to identify more subtle movement patterns from accelerometry by taking into consideration the full 3-axis, high-frequency data before and after the time period that is labelled. This provides the following measures: 

  • Global – sleep efficiency over the whole night
  • Onset –sleep onset latency
  • Disruption – number of awakenings per night

The DLS algorithm has been developed on data acquired from subjects with different neurodegenerative diseases and is currently further developed and validated in several clinical studies.