Making cognitive therapeutic care accessible and personalised for populations at-risk of cognitive decline and patients diagnosed with early to moderate-stage dementia*
* To date, Lilia is in the research and development phase and has not yet been the subject of scientific publications. It cannot be considered a medical device at this stage.
Lilia is an Artificial Intelligence Therapeutic
How it works?
Involving the family to better understand the patient's biography and personality.
Analyse the history of interactions with the patient over the sessions and model motivations, interests, real disorders, & and cognitive abilities.
Receive insights from caregivers and therapists regarding the manifestation of certain cognitive deficits.
Encourage the patient to interact with her and offer meaningful content and activities
Adapt to the patient's changing capacities (attention, memory, concentration, etc.)
Cognition and patient health status evolve; the artificial intelligence Lilia too.
Lilia will take into account the patient's biography and all information provided by relatives and caregivers to fully personalize its interactions and the content it offers. Dynamically, Lilia constantly seeks to:
● Adapt content to the patient's faculties
● Propose content that stimulates the patient's cognition the most
The strategy implemented by Lilia combines pedagogical content inspired by different cognitive therapies in a hybrid approach:
Reminiscence, involving initiated discussions about past memories and experiences
Cognitive Stimulation, aiming to bring out new ideas, maximize preserved cognitive abilities, and express opinions rather than facts
Art therapy, which uses art as a psychotherapeutic means to help patients express a personal vision while engaging in creation
Music therapy, aiming to facilitate and promote communication and social relations through listening to music and producing sounds
The artificial intelligence Lilia will seek continuous training by modeling the clinical impacts its interactions with the patient might produce (evidence-based feedback loop). Research by KompanionCare aims to understand the extent to which patient interactions with Lilia could constitute self-questionnaires generated by the patient, allowing tracking of cognitive impairment: it is necessary to measure the correlation level of responses to these self-questionnaires with evaluation grids usually used by therapists to assess disease progression (MMSE and ACE-III and NPI).
Hybrid Cognitive Approach:
The Role of the Feedback Mechanism:
The Reasons Behind Developing Lilia:
A very large number of people are affected by dementia and too few neurologists to take care of them
2,980 neurologists must take care of more than one million patients in France, with 220,000 new cases per year.
The Importance of Cognitive Therapies in Disease Progression
To date, non-pharmaceutical therapies based on cognitive stimulation have demonstrated an improvement in cognitive functions and patient quality of life comparable to pharmacological approaches (*). It is interesting to note that group cognitive stimulation is the only non-pharmacological intervention recommended by the UK's National Institute for Clinical and Health Excellence for people with dementia.
* Reference: Spector A. Thorgrimsen L. Woods B.O.B. Royan L. Davies S. Butterworth M. Efficacy of an evidence-based cognitive stimulation therapy programme for people with dementia: randomized controlled trial. Br J Psychiatry. 2003;183:248–254
The Need for a Personalised Approach
Two main directions must be taken into account in cognitive therapy:
Individual and/or group sessions with the patient on themes and situations that will allow them to use the language acts they can still emit correctly: the more these acts are used by the patient, the longer they will remain within their competence.
Taking into account and adapting the environment (caregivers, assistants, family, etc.) to the changes in the patient's communication capacities.
Taking into account the Isolation Factor of Patients in the Evolution of the Disease.
Isolation is a cause and a symptom.
The Difficulty of Making Cognitive Therapy Accessible to All Patients
Studies show that cognitive therapies are both effective and cost-effective, but their current availability is limited and access is reserved for a minority.*
*Knapp M, Bauer A, Wittenberg R, et al. What are the current and projected future cost and health-related quality of life implications of scaling up cognitive stimulation therapy? Int J Geriatr Psychiatry. 2021;1‐10. doi:10.1002/gps.5633
The stakes of the research and projects led by
Isolation is both a cause and a symptom of dementia disease, increasing the risk of depression, anxiety, and violent behaviors. Lilia could help reduce patients' feeling of isolation by:
Improving patient follow-up by the medical profession and helping them feel less alone and neglected in managing the disease and side effects related to potential treatments.
Interacting regularly with the patient and encouraging them to communicate with their loved ones.
The development of Lilia takes into account these two dimensions in the construction of interactions with the patient: Lilia will constantly seek to modulate its communication behavior to adapt to that of the patient.
Several scientific studies have also established a link between the lack of cognitive stimulation and an earlier onset of the first signs of dementia: these studies show that cognitive stimulations delay the first signs of dementia disease by an average of five years.
Cognitive stimulation thus constitutes a major challenge for prevention but also for slowing down the progression of the disease.
New patient monitoring tools need to be developed to allow practitioners to follow a larger number of patients while maintaining the quality of their care.
Kompanion Care is to determine to what extent recent advances in digital technologies can democratise patients' access to cognitive therapies:
Due to their reduced cost
The acceptability of patients to receive cognitive therapy in a semi-autonomous manner
Digital Technologies & Cognitive Therapies
Kompanion Care's research aims to evaluate the effectiveness of a digital cognitive therapy based on frequent interaction (at least twice a week for more than 4 months) of an intelligent digital conversational agent with patients suffering from dementia with mild to moderate cognitive impairment.
The specificity of the approach lies in leveraging major advances in the last two years in Natural Language Processing (NLP) and more broadly in artificial intelligence to ultimately offer patients a hybrid cognitive therapy based mainly on cognitive stimulation, reminiscence, and musical & artistic cognitive therapies.
General Research Context
Demonstrated Effectiveness of Non-Pharmaceutical Treatments, Particularly Cognitive Stimulation Therapies
Dementia affects over one million people in France and over 55 million worldwide, with a rapidly increasing incidence rate, particularly due to population aging. The disease is characterized by cognitive impairments and psychological and behavioral symptoms (neuropsychiatric symptoms).
Non-pharmaceutical therapies based on cognitive stimulation have demonstrated improvements in cognitive functions and patient quality of life, comparable to pharmacological approaches  and . Furthermore, group cognitive stimulation is the only non-pharmacological intervention recommended by the UK's National Institute for Clinical and Health Excellence for people with dementia.
Evidence of the Contribution of Specific Digital Tools for Implementing Cognitive Stimulation Therapies
Previous research and clinical studies have been conducted to quantify the impact of using digital tools (Computer Interactive Reminiscence and Conversation Aid -- CIRCA, Tablet Health Care Platform -- ehcoBUTLER, Memory Motivation -- MeMo, X-Torp, etc.) by dementia patients. These tools aim to encourage patients to communicate, make independent choices, work on long-term memory, minimize working memory disorders, and improve their quality of life and propensity to participate in group activities.
These studies have shown significant improvements:
For patients: evidence of the positive impact of using these tools on the cognitive state, behavior, or quality of life of patients .
For caregivers: increased feeling of competence in providing care  and changing their perception of the patient relationship .
However, using digital tools by individuals with cognitive and behavioral disorders faces challenges in tool handling and motivation to use it.
Significant progress has been made in Natural Language Processing research since 2020, enabling the construction of conversational artificial intelligences producing content indistinguishable from human-written content in certain usage contexts . These recent advances open new research perspectives in developing cognitive stimulation therapies. Kompanion Care's exploration and research continue these works.
Les recherches lancées et futures de Kompanion Care poursuivent principalement les objectifs suivants :
Kompanion Care's ongoing and future research primarily aims to:
Evaluate the extent to which data generated by patient interactions with Livio is representative of the patient's level of cognitive impairment, quality of life, and overall health.
Assess the patient's acceptability of the digital conversational agent, their propensity to want to use the tool spontaneously, the limitations for the patient to use it autonomously, and evaluate how technology could make cognitive stimulation therapies more accessible to a larger number of patients at a competitive cost.
Quantify the clinical impact of regular interactions of the intelligent conversational agent with the patient (cognition, general health level, quality of life). Evaluate to what extent the use of the tool could delay the patient's departure to a specialized center.
Measure the clinical impact of personalized patient care by therapists resulting from continuous monitoring of the patient's cognitive state by the digital tool.
Measure the impacts on the quality of life of caregivers and the quality of their relationship with the patients resulting from the use of the tool.
Launching 2 Research Projects
Kompanion Care has initiated two research projects launched in 2022, drawing from interactions between patients and the artificial intelligence Lilia.
Lilia e-Pro Project
Lilia Therapy Project
In many therapeutic areas (oncology, cardiology, etc.), telemonitoring has demonstrated its positive impact on the overall survival of patients. In the case of dementia, data collection via an e-PRO system (electronic patient-reported), where patient interactions auto-generate follow-up data/markers of cognitive faculties, does not exist to our knowledge. Self-questionnaires enable direct collection of information on disease progression. The analysis of these questionnaires enriches clinical reality with practical reality.
Livio e-Pro research aims to evaluate the extent to which data generated by patient interactions with Livio is representative of the patient's level of cognitive impairment. The correlation between auto-generated patient data and their responses to MMSE, ACE-III, CDR, and NPI evaluation tests (4 evaluation grids predominantly used in the literature) will be evaluated. Due to their frequency, granularity, and coverage of many aspects of cognition, the quantitative data generated by tool use will be studied to enrich dementia state evaluation grids with conversational quantitative measures (e.g., speech duration, frequency of statements) of socialization level, orientation, apathy, and attention.
Group cognitive stimulation therapies have demonstrated their effectiveness in cognitive and patient quality of life . However, numerous studies on various cognitive therapies (reminiscence therapy, music therapy, cognitive training therapy, etc.) provide evidence of their effectiveness on specific aspects of cognition or patient health state, but not always consistently.
Furthermore, it has been demonstrated that patient-centered care interventions reduce agitation, neuropsychiatric symptoms, and depression, and improve quality of life, confirming the interest in developing personalized therapies.
The research focuses on the effectiveness that could represent the hybridization of 4 personalized cognitive therapies: i. cognitive stimulation, ii. reminiscence, iii. art therapy, and iv. music therapy.
Dynamic content selection drawn from different therapies is performed by the artificial intelligence Livio, which seeks to maximize its clinical effectiveness by observing the effects of its interactions with the patient. The artificial intelligence uses self-generated patient evaluations (see the research project section on Livio e-Pro) to assess its impacts and adapt the behavior of its conversational agent.
The quantitative objectives of this therapeutic approach are to measure:
The acceptability of treatment by the patient and their ability to receive it autonomously
The impact of treatment (depending on dosage - duration and frequency of administration) on cognition, quality of life, and patient health state.
 Spector A., Thorgrimsen L., Woods B.O.B., Royan L., Davies S., Butterworth M. Efficacy of an evidence-based cognitive stimulation therapy programme for people with dementia: randomised controlled trial. Br J Psychiatry. 2003;183:248–254
 Gibbor L, Yates L, Volkmer A, Spector A. Cognitive stimulation therapy (CST) for dementia: a systematic review of qualitative research. Aging Ment Health. 2021 Jun;25(6):980-990. doi: 10.1080/13607863.2020.1746741. Epub 2020 Apr 7. PMID: 32252561.)
 Astell A.J., Smith S.K., Potter S., Preston-Jones E. Computer Interactive Reminiscence and Conversation Aid groups—Delivering cognitive stimulation with technology. Alzheimer’s Dement. Transl. Res. Clin. Interv. 2018;4:481–487. doi: 10.1016/j.trci.2018.08.003
 Contreras-Somoza LM, Toribio-Guzmán JM, Sánchez-Gómez MC, Irazoki E, Martín-Cilleros MV, Verdugo-Castro S, Parra-Vidales E, Perea-Bartolomé MV, Franco-Martín MÁ. Perceptions of Older People with Cognitive Impairment, Caregivers, and Professionals about ehcoBUTLER (Tablet Health Care Platform): A Qualitative Focus Group Study. Int J Environ Res Public Health. 2022 Jun 1;19(11):6761. doi: 10.3390/ijerph19116761. PMID: 35682344; PMCID: PMC9180242.
 Philippe Robert, Valeria Manera, Alexandre Derreumaux, Marion Ferrandez y Montesino, Elsa Leone,
et al.. Efficacy of a Web App for Cognitive Training (MeMo) Regarding Cognitive and Behavioral Per-
formance in People With Neurocognitive Disorders: Randomized Controlled Trial. Journal of Medical
Internet Research, JMIR Publications, 2020, 22 (3), pp.e17167. 10.2196/17167
Ben-Sadoun G, Sacco G, Manera V, Bourgeois J, König A, Foulon P, et al. Physical and Cognitive Stimulation Using an Exergame in Subjects with Normal Aging, Mild and Moderate Cognitive Impairment. Journal of Alzheimer’s Disease. 2016;53:1299-314
 Astell A.J., Ellis M., Bernardi L., Alm N., Dye R., Gowans G. 22nd Conference of Alzheimer's Disease International. 2007. Developing technology to support the relationship between people with dementia and caregivers
 Astell A.J., Alm N., Gowans G., Ellis M., Dye R., Vaughan P. Involving older people with dementia and their carers in designing computer based support systems: some methodological considerations. Universal Access Inf Soc. 2009;8:49
 Brown T., Mann B., Ryder N., Subbiah M., Kaplan J. D., Dhariwal P., ... & Amodei D. (2020). Language models are few-shot learners. Advances in neural information processing systems, 33, 1877-1901
 Sun Y, Zhang X, Wang Z. Comparative Effectiveness of 3 Settings of Cognitive Stimulation Therapy on Cognition and Quality of Life for People With Dementia: A Systematic Review and Network Meta-analysis. J Am Med Dir Assoc. 2022 Mar;23(3):461-467.e11. doi: 10.1016/j.jamda.2021.11.015. Epub 2021 Dec 20. PMID: 34942157
 Kim SK, Park M. Effectiveness of person-centered care on people with dementia: a systematic review and meta-analysis. Clin Interv Aging. 2017 Feb 17;12:381-397. doi: 10.2147/CIA.S117637. PMID: 28255234; PMCID: PMC5322939.