Organizations: The Chinese University of Hong Kong · The University of Hong Kong · Guangdong Laboratory of Artificial Intelligence and Digital Economy (SZ) · Huawei
Cognitive impairment is becoming a major public health challenge. Cognitive Stimulation Therapy (CST) is an effective intervention for cognitive impairment, but traditional methods are difficult to scale, and existing digital systems struggle with group dialogues and cognitive stimulation principles. While Large Language Models (LLMs) are powerful, their application in this context faces key challenges: cognitive stimulation dialogue paradigms, a lack of therapeutic reasoning, and static-only user modeling. To address these issues, we propose a principle-driven adaptive policy actualized through a Group Cognitive Stimulation Dialogue (GCSD) system. We first construct a dataset with over 500 hours of real-world CST conversations and 10,000+ simulated dialogues generated via our Principle-Guided Scenario Simulation strategy. Our GCSD system then integrates four core modules to overcome LLM limitations: (i) a multi-speaker context controller to resolve role confusion; (ii) dynamic participant cognitive state modeling for personalized interaction; (iii) a cognitive stimulation-focused attention loss to instill cognitive stimulation reasoning; and (iv) a multi-dimensional reward strategy to enhance response value. Experimental results demonstrate that GCSD significantly outperforms baseline models across various evaluation metrics. Future work will focus on long-term clinical validation to bridge the gap between computational performance and clinical efficacy.
Figures & tables
Figure 1: The construction of real and simulated datasets.
Categories
Num. (Real)
Num. (Stimulation)
Training Dataset
Data Count
10,502
9,429
Total Tokens
407,699
295,902
Assistant Count
160,619
119,370
Assistant Tokens
1,644
119,511
Human/Elder Number
9
7
Table 1: Statistics of the multi-party cognitive stimulation dialogue dataset.
Figure 2: This is an overview of the GCSD. On the left side is the data construction phase, which includes the workflows for both real data and simulated data construction (principle-guided scenario simulation strategy), as well as an example shown at the lower left. On the right side is the GCSD model phase, where the model first learns patterns and scenarios based on simulated data, and then trains on real data. The model consists of four parts: multi-speaker context controller, dynamic participant cognitive state modeling, cognitive stimulation-focused attention Loss, and multi-reward policy optimization.
Models
ROUGE-L
BLEU-2
BLEU-4
BERTScore
Distinct-2
Relevance
Empathy
Fluency
Simulated Dataset
ERNIE
17.24
8.32
7.38
65.35
53.92
3.05
3.20
2.90
Doubao-Pro
23.58
22.39
21.51
73.19
71.03
4.00
3.55
3.33
GPT-4o
21.78
20.38
20.05
72.46
67.83
3.67
3.65
3.45
DeepSeek-671B
21.67
22.13
22.09
77.29
73.79
3.59
3.45
3.36
LLaMA-3.1-405B
22.56
22.13
21.08
76.52
71.26
3.75
3.70
3.42
Table 2: Evaluation results between baselines and our GCSD on the cognitive stimulation simulated and real testing datset . The first four metrics are automatic metrics, while the last three metrics are human metrics. Bold denotes the best results.
Simulating Standardized Patients with cognitive impairment offers a scalable and ethical solution for clinical training. However, existing methods rely on discrete prompt engineering and fail to capture the heterogeneity of deficits across varying domains and severity levels. To address this limitation, we propose StsPatient for the fine-grained simulation of cognitively impaired patients. We innovatively capture domain-specific features by extracting steering vectors from contrastive pairs of instructions and responses. Furthermore, we introduce a Stochastic Token Modulation (STM) mechanism to regulate the intervention probability. STM enables precise control over impairment severity while mitigating the instability of conventional vector methods. Comprehensive experiments demonstrate that StsPatient significantly outperforms baselines in both clinical authenticity and severity controllability.
Weikang Zhang, Zimo Zhu, Zhichuan Yang +3
School of Software, Xi’an Jiaotong University · Institute of Data Science, National University of Singapore · College of Computer Science, Sichuan University +1
Emotional Support Conversation (ESC) plays a critical role in mental health assistance by providing accessible psychological support in real-world applications. Large Language Models (LLMs) have shown strong empathetic abilities in ESC tasks. Yet, existing methods overlook the issue of cognitive distortions in help-seekers' expressions. As a result, current models can only provide basic emotional comfort, rather than helping help-seekers address their psychological distress at a deeper cognitive level. To address this challenge, we construct the CogBiasESC dataset, the first dataset that expands existing ESC datasets by adding labels for cognitive distortions, includes their type, intensity, and safe risk level. Furthermore, we propose the Cognitive Policy-driven Large Language Model framework (CoPoLLM) to enhance LLMs' ability to diagnose and intervene cognitive distortions in help-seekers. We also analyze the safety advantages of CoPoLLM from a theoretical perspective. Experimental results show that CoPoLLM significantly outperforms 15 state-of-the-art baselines in terms of distortion diagnosis accuracy, intervention strategy effectiveness, and safety risk control.
Lin Zhong, Renjin Zhu, Shujuan Ma +4
Harbin Institute of Technology, Shenzhen, China · City University of Macau, Macao SAR, China · Peng Cheng Laboratory, Shenzhen, China
The application of games as a therapeutic tool for cognitive training is beneficial for patients with cognitive impairments. However, effective game design for individual patient is resource-intensive. To this end, we propose an LLM-powered method, \ours, for automated and personalized therapeutic game design. Inspired by the Dungeons & Dragons, LETGAMES generates an open-world interactive narrative game. It not only generates game scenarios and challenges that target specific cognitive domains, but also employs conversational strategies to offer guidance and companionship. To validate its efficacy, we pioneer a psychology-grounded evaluation protocol LETGAMESEVAL, establishing comprehensive metrics for rehabilitative assessment. Building upon this, our experimental results from both LLM-based assessors and human expert evaluations demonstrate the significant potential of our approach, positioning LETGAMES as a promising solution to the widespread need for more accessible and tailored cognitive training tools. Our code will be open-sourced upon acceptance.
Jingwei Shi, Shengyu Tao, Xinxiang Yin +3
Shanghai University of Finance and Economics, China · Northwest Polytechnical University Xi’an, China · Institute of Data Science, National University of Singapore, Singapore +1