Enhancing Virtual Avatar Performance with CTA