This thesis investigates the development and testing of a novel tablet-based speech comprehension therapy, for persons with chronic aphasia (PWA). In Chapter 1 of this thesis, I present a qualitative study on the development of a speech comprehension therapy app (Listen-In), with gamification. Five co-design focus groups were carried out, resulting in an iterative cycle of prototype development. Using thematic analysis, a number of barriers to usability and enjoyability were identified. This resulted in a multitude of design changes, which led to a final product suitable to be self-administered by patients. In Chapter 2, I analysed data from a cross-over clinical trial of Listen-In (N=35), which compared 12-weeks of Listen-In treatment with 12-weeks of standard care. These findings showed: (i) PWA made large and significant gains in speech comprehension for words trained during treatment; (ii) gains were item specific, suggesting facilitation of item-specific networks; (iii) a combination of baseline measures explained only a small amount of variability in treatment outcomes. In Chapter 3, I used voxel-based morphometry (VBM) to investigate whether baseline structural integrity contributed to treatment outcomes, in a subgroup of patients (N=25). I found that greater volume of mainly white matter, in distributed regions in the right hemisphere, predicted greater response to treatment. I speculated that these regions related to multi-functional networks supporting a range of cognitive and language functions. In Chapter 4, I used longitudinal VBM to investigate therapy-driven structural neuroplasticity in the same subgroup of patients. I found that greater improvements in comprehension of treated items, were related to tissue changes in bilateral temporal lobes, in key speech processing regions. In the final discussion section, I interpret these key findings, their clinical implications, and directions for future research.