The conversational generation of business process models with LLM-based chatbots offers promising perspectives for involving non-experts in process modeling. However, corresponding approaches will not be successful if the created models are of bad quality and, are therefore not comprehensible. While cognitive factors play a crucial role in the comprehension of process models, research on the specific factors that foster or affect the comprehensibility of process models created by humans, LLM-based bots, or a combination of them remains limited. This PhD research aims to systematically investigate the cognitive factors (e.g. perception, attention, memory, problem-solving, and metacognition) relevant in this context. Following the Design Science Research Methodology, this PhD research develops the conceptual framework ComprehenGen, which shall integrate the cognitive factors to systematically assess and provide guidelines for enhancing the comprehensibility of LLM-generated process models. The evaluation of this framework will incorporate methods from both cognitive neuroscience (e.g., eye tracking) and cognitive psychology (e.g., Cognitive Load Theory) to enable a comprehensive understanding of how humans process and interpret these models. Ultimately, this PhD research seeks to provide a structured approach to improve the comprehensibility of LLM-generated process models and their practical applicability in business process management.