In the present invention a content distribution network delivers content that can include training, entertainment, assessment, and evaluation among many other types of content. Content can be delivered on-demand to devices operated by users in local or remote locations or can be delivered live in a present local or live-streamed to remote locations. In each of these cases the user, thru the user device, can provide valuable feedback to the content distribution network to improve content delivery and user content retention in the form of correlating detected patterns with the content being delivered. An image sensor attached to a user device captures facial expressions as patterns and transmits them to a content manager that aligns the pattern with the content and determines a sentiment corresponding to the pattern. The content manager can improve the user's content reception by implementing a workflow responsive to the corresponding sentiment pattern.