Evaluating usability and acceptance of a socially assistive robot supported cognitive training for depression – results of the randomized controlled pilot study ‘AMIGA’

The integration of socially assistive robots (SARs) in mental healthcare offers promising opportunities to enhance traditional treatments. The humanoid SAR “Pepper” has shown potential for supporting cognitive and emotional interventions for individuals. While SAR acceptance has been explored in general healthcare, limited evidence exists regarding its usability and acceptance among individuals with affective disorders. This study aimed to assess the usability and acceptance of “Pepper” as an adjunct motivational technology in combination with tablet-based cognitive training, compared to tablet training alone, among inpatients with depression, focusing on sex-specific differences. A randomized controlled trial was conducted between June and October 2024 with 32 inpatients diagnosed with depression. Participants were randomly assigned to one of two groups: the SAR group, which used “Pepper” together with the cognitive training app “Multimodal Activation” in combination with motivational feedback, and the control group, which used the same app on a tablet without motivational feedback. Each participant completed two cognitive training sessions of approximately 10 to 20 minutes. Standardised questionnaires, namely, Technology Acceptance Model (TAM), System Usability Scale (SUS), and Technology Usage Inventory (TUI), were administered after the second session. To analyze group and sex differences, analyses of co-variance were used for single-scale measures (SUS, TUI Overall), and multiple analyses of co-variances were used for instruments with multiple subscales (TAM, TUI). A significant group difference was found in the SUS score, favoring the SAR group. The SAR group also scored higher in the TUI User-Friendliness, while the tablet group showed higher scores in Accessibility. Regarding sex differences, female participants scored higher on the TUI Overall and Scepticism subscales than male participants. These findings suggest that SAR-supported cognitive training is a viable and accepted tool for supporting cognitive training in psychiatric care. High usability and positive acceptance ratings indicate its potential to complement conventional therapies. The observed sex-specific differences underline the relevance of tailored robotic interventions. Limitations of this study include the small sample size and short intervention period. Further studies are warranted to validate these findings and examine the ethical considerations and human–robot interaction dynamics in psychiatric settings.

1 Introduction

Depression is a widespread mental health disorder that affects over 264 million individuals worldwide (1). Treating depression often requires a multidimensional approach, including pharmacological treatment, psychotherapy/psychosocial interventions, and social support (2). In this context, socially assistive robots (SARs) could represent an innovative complement to traditional therapies and interventions. SARs are designed to interact with humans and assist them in various domains (3). The integration of robotic technology into healthcare has gained significant importance in recent years, particularly in mental health (4). SARs, such as the humanoid robot “Pepper”, offer promising opportunities to support patients with mental health conditions, particularly within clinical and therapeutic settings. By enhancing social interaction during therapy sessions, promoting cognitive stimulation through interactive exercises, and fostering emotional well-being in individuals with depression and anxiety disorders, SARs can contribute significantly to mental healthcare. These effects are facilitated by socially assistive behaviors, including empathetic verbal responses, adaptive nonverbal communication (e.g. gestures, gaze, or posture), and the ability to tailor interactions based on the user’s emotional state (5–7). However, the usability and acceptance of this technology among psychiatric patients, particularly those with depression, remain underexplored, and it is still unclear which patient groups are particularly responsive to SARs interventions, and which may find them less suitable.

Compared to other digital intervention platforms such as tablets or smartphones, SARs offer distinct advantages in promoting engagement and adherence in mental health interventions. Their embodied, human-like form enables multimodal communication, combining speech, gestures, and facial expressions, which fosters a stronger sense of social presence and emotional connection with users (3, 5, 6). Previous studies have shown that embodied agents can enhance motivation and therapeutic alliance more effectively than screen-based devices, particularly among individuals with affective or cognitive impairments (8–10).

While tablet-based cognitive training provides valuable accessibility and familiarity, it remains limited to two-dimensional interaction and lacks reciprocal social feedback. In contrast, humanoid SARs can actively perceive and respond to users’ emotions, deliver verbal encouragement, and adapt their behavior in real time. These features make them particularly suitable for psychiatric rehabilitation, where emotional engagement and motivational support are critical for therapy adherence (11, 12).

The usability and acceptance of SARs are essential factors for their successful integration into clinical settings (13). Attitudes towards robots are influenced by cultural, demographic, and individual factors (14). High usability is crucial for both patients and healthcare providers because it facilitates effective use and encourages acceptance (15). This is especially important for patients with depression, where cognitive impairments and emotional factors may pose additional challenges to usability (16). Acceptance refers to patients’ willingness to interact with the SAR and incorporate it into their treatment, whereas usability describes the robot’s user-friendliness and effectiveness in practical applications (3). These aspects are particularly significant in healthcare, where perceived usefulness, user-friendliness, and trust in technology play pivotal roles (13). Research has consistently shown that the overall usability and acceptance of robots in healthcare are positive (5, 17–20). The usability and acceptance of SARs vary among different patient groups. For example, Pu et al. (21) found that older adults generally have a positive attitude towards SARs, particularly when these robots are perceived as supportive rather than replacements for human interaction. In psychiatric care, Huijnen et al. (22) observed that patients with autism spectrum disorders benefit from interacting with SARs, indicating the potential usability and acceptance of SARs within this group. Broadbent et al. (3) emphasized that aspects of user-friendliness, such as intuitive interaction design, adaptability to user needs, and smooth navigation with minimal technical errors, are critical determinants of acceptance and willingness to engage with healthcare robots among older adults. This underscores the importance of designing intuitive and accessible human-robot interactions to ensure the successful adoption of SAR in diverse clinical settings.

Furthermore, baseline clinical factors such as depression severity may influence patients’ engagement with digital or robot-assisted interventions. Higher symptom severity has been associated with reduced motivation, concentration, and psychomotor activity, which can limit active participation and perceived usability (23, 24). Likewise, pharmacological treatment may affect cognitive processing speed or emotional responsiveness, potentially moderating interaction quality and engagement. Considering these aspects is essential for understanding variability in patient responses to SAR-supported interventions (25).

Sex-specific differences in the acceptance and use of technology are also important aspects that must be considered in SAR research, particularly in psychiatric settings. Previous studies suggest that men and women may exhibit different preferences and attitudes towards robots, with some findings indicating higher acceptance of caregiving robots among women (26–28). These differences may significantly influence the perception and use of robots in clinical contexts. Despite the growing interest in SAR in healthcare, there is a lack of studies focusing on the usability and acceptance of such technologies among psychiatric inpatients with depression. Existing research has primarily examined older adults, individuals with dementia, and general medical populations (13, 14). Thus, there remains a need to explore how psychiatric patients, particularly those with depression, experience and evaluate SAR, while also accounting for sex-specific perspectives to inform the development of more tailored and sensitive robotic interventions (26).

This study aims to address recent research gaps by investigating the usability and acceptance of the SAR “Pepper” when used as an adjunct motivational technology in combination with tablet-based cognitive training, compared with tablet-based training alone without motivational feedback, among inpatients with depression in psychiatric care. Specifically, this study investigated how inpatients with depression in psychiatric care evaluate the usability and acceptance of a SAR-enhanced intervention, while additionally examining sex-specific differences in patients’ evaluation of the intervention. For this purpose, we hypothesized that patients with depression who complete cognitive training sessions using the SAR “Pepper” with motivational feedback would report higher 1) usability and 2) acceptance than those receiving the same training via tablet, with significant sex-specific differences in their evaluations.

2 Methods and materials

2.1 Design, setting and sample

This pilot study was designed and conducted as a randomized controlled parallel two-arm trial (RCT) following the recommendations of the CONSORT 2025 Statement (29) to ensure transparency and methodological rigor in reporting. The RCT was carried out between June and October 2024 at the Clinical Division of Psychiatry and Psychotherapeutic Medicine at the Medical University of Graz, Austria. The sample consisted of 32 inpatients with depression. The diagnosis of depression was made by the treating clinicians according to the ICD-10 criteria. To evaluate the severity of the current depressive symptomatology the Montgomery-Åsberg Depression Rating Scale (MADRS) (30, 31) and the Beck Depression Inventory – Revised (BDI-II) (32) were used. The MADRS is an observer-rated assessment tool that evaluates depressive symptoms over the past week. It consists of ten items rated on a 7-point Likert scale, with total scores ranging from 0 to 60. Scores between 0 and 6 indicate no depression, whereas scores above 34 suggest severe depression (30, 31). In contrast, the BDI-II is a self-report questionnaire in which patients assess their feelings during the past week. It comprises 21 items rated on a 4-point Likert scale, yielding a maximum score of 64 points. A score of 0–8 points indicates no signs of depression, 9–13 points suggests minimal depression, 14–19 points indicates mild depression, 20–28 points indicates moderate depression, and scores of 29 or higher indicate severe depression (32). All assessments were conducted by trained staff members in a designated quiet room that was specifically prepared for this purpose. Both tests were used solely to describe the basic characteristics of the sample and were not considered as outcome variables.

2.2 Sample size

The sample size calculation was originally based on a repeated-measures ANCOVA design with two measurement points (η² = 0.3, power = .90, α = .05), yielding a target of N = 32 participants (33). This approach was chosen because the broader project “AMIGA” (34, 35) included multiple pre–post measures, which are not part of the current publication. Although the present analysis focuses on post-intervention questionnaire data (SUS, TAM, TUI), the same total sample size was retained, as it meets the requirements for parametric testing based on the central limit theorem, which is generally satisfied for sample sizes of approximately 30 or more. “AMIGA” was conceptualized as a pilot study intended to generate initial evidence on usability and acceptance of SAR-supported cognitive training in depression. Consequently, the findings should be interpreted as exploratory and hypothesis-generating rather than confirmatory. The calculation was conducted using G*Power software (version 3.1.9.6) (36, 37).

2.3 Inclusion and exclusion criteria

The inclusion and exclusion criteria for the participants were as follows. Participants eligible for inclusion were men and women aged ≥ 18 years who met the ICD-10 criteria for moderate or severe depression. They had to be able to speak and understand German, give written informed consent, and had no physical limitations that would prevent the use of the SAR “Pepper”.

The exclusion criteria were refusal to participate, presence of psychotic disorders, organic brain diseases or dementia, substance-induced disorders, or acute suicidality.

2.4 Randomization and blinding

After the participants agreed to enroll, allocation to the intervention or control group was carried out through randomization using an online randomizer (https://www.studyrandomizer.com) (38), ensuring balanced randomization with respect to 1) sex, 2) depression severity according to the ICD-10 diagnosis at admission, and 3) age. Blinding was not performed in this study because the interventions were apparent and could not be concealed.

2.5 Study course

First, the participants were assigned to either an intervention group (SAR “Pepper” with motivational feedback) or a control group (tablet without motivational feedback). Over five days, participants in both groups had the opportunity to engage in cognitive training using the “Multimodal Activation” (MMA) application twice. In the intervention group, SAR “Pepper” was used to accompany and monitor the participants during training. After the participants completed the three tasks in cognitive training, SAR “Pepper” provided motivational feedback to encourage their progress. In contrast, the control group received the same cognitive training exercises, but these were conducted exclusively using the “MMA” app on an Android tablet (see Figures 1, 2). Apart from this difference, the procedures were identical in both groups. The primary distinctions between the intervention and control conditions were the mode of delivery (SAR “Pepper” versus tablet) and the presence or absence of motivational feedback during training (no motivational feedback in the tablet group).

Figure 1

Figure 1. Study course and patient flow diagram according to Hopewell et al. (29).

Figure 2

Figure 2. Participant interacts with the tablet-supported MMA application on the SAR “Pepper” and completes the cognitive task (solving arithmetic problems) (The participant provided written informed consent for this photograph to be taken and for its publication in the article; Credits: Sandra Draxler – JOANNEUM RESEARCH DIGITAL).

The first training session began with the participants completing the baseline questionnaires (t₀) via LimeSurvey (https://www.limesurvey.org) (39). They then participated in the first cognitive training session. In the second session, they again completed a cognitive training session and filled out the post-intervention questionnaires (t₁) using LimeSurvey. Subsequently, the participants completed the study. Each cognitive training lasted approximately 10 to 20 minutes.

2.6 SAR “Pepper”

SAR “Pepper” is a 120 cm tall and 28 kg heavy humanoid robot. SAR “Pepper” is capable of moving its arms, hands, and head and can analyze and respond to human emotions through body posture, speech, facial expressions, and voice tone. Equipped with microphones, SAR “Pepper” can identify and turn towards the person speaking (40–42). SAR “Pepper” fosters friendly and effective human-robot interaction to support therapy, rehabilitation, or learning, often enhancing motivation and quality of life (43).

2.7 “MMA”-app and cognitive training

The “MMA” (m_ultim_odal a_ctivation) application (19, 44) consists of a tablet-based front end and a back end installed on a central server. Both components were developed using Microsoft’s “.NET Framework”. The front end was created using Xamarin, a cross-platform framework by Microsoft that enables mobile application development in “.NET”. The application was specifically designed for Android and optimized for tablet displays. Additionally, an application programming interface was implemented to facilitate user management and data exchange, including authentication, version updates, and the transmission of exercise results. A structured query language database was developed to store the user data and the received results. Furthermore, a content management system was implemented, allowing the creation and categorization of various types of exercises into thematic units (45, 46).

Cognitive training consisted of three different exercise blocks, each comprising six tasks. Consequently, a total of 18 tasks were completed per training session. The difficulty level was adjusted after the completion of each exercise block and progressively increased with each step. The first exercise block was conducted at difficulty level 2 out of 4, the second block at difficulty level 3 out of 4, and the third block at the highest difficulty level, 4 out of 4.

Each training session included only three different cognitive tasks selected from distinct cognitive domains. In the first training session, participants were required to complete tasks such as finding picture pairs (domain: memory), solving arithmetic problems (domain: working memory), and filling in missing words in a text (domain: semantic memory). In the second training session, the tasks consisted of solving arithmetic problems (domain: working memory), completing puzzles (domain: visuospatial memory), and organizing a sequence of events in the correct order (domain: seriality and long-term memory).

The selection of cognitive tasks was based on established domains known to be affected in cognitive training, including memory, attention, executive functioning, and verbal fluency. Tasks were derived from the validated MMA training framework, which has previously been used in studies with neurological and psychiatric populations (19, 44–46). The difficulty levels were designed to progressively increase across sessions to sustain motivation and allow participants to experience a sense of achievement, while avoiding cognitive overload. This adaptive structure follows cognitive rehabilitation principles suggesting that gradual task escalation enhances engagement and supports learning in individuals with cognitive and affective impairments (47).

2.8 Questionnaires

2.8.1 System usability scale

The System Usability Scale (SUS) is a questionnaire designed to evaluate technology usability. The advantage of the SUS lies in its technology-independent nature and brevity, making it quick and easy to complete. It consists of 10 questions rated on a five-point Likert scale (0 = strongly disagree to 4 = strongly agree). The items (2, 4, 6, 8, 10) need to be reverse-scored before further calculations, as they are negatively worded. Subsequently, the sum of all 10 items was calculated and multiplied by 2.5. This step yielded a total score (SUS score). A total score of 68 or higher indicates good usability (48–50). In the literature, the SUS is described as a reliable scale (ω = .89 and Cronbach’s alpha α = .86) that effectively detects differences compared to other commercially available questionnaires, even with small sample sizes (51). Its validity, in this case, perceived user-friendliness, was also rated highly (52).

2.8.2 Technology usage inventory

The Technology Usage Inventory (TUI) evaluates the psychological and technology-specific factors that influence technology adoption. The TUI instrument consists of 30 items distributed across eight subscales: Curiosity, Anxiety, Interest, User-Friendliness, Usefulness, Scepticism, and Accessibility. The Immersion subscale, although part of the standard TUI, was intentionally excluded from this study, as immersion was not applicable to either the robot or tablet condition. This modification aligns with the TUI manual guidelines. Each item was rated on a seven-point Likert scale (1 = “Strongly Disagree” to 7 = “Strongly Agree”). Subscales with three items (Accessibility and User-Friendliness) yielded scores ranging from 3 to 21 points, while subscales with four items (Anxiety, Curiosity, Interest, Usefulness, and Scepticism) ranged from 4 to 28 points, resulting in a total possible TUI score between 26 and 182 points. Additionally, the Intention to Use scale, comprising three items, was included, allowing participants to rate their intentions on a scale from 0 to 100 per item, for a total score ranging from 0 to 300. The TUI has demonstrated robust psychometric properties, with internal consistencies (Cronbach’s alpha) ranging from.70 to.89 (53).

2.8.3 Technology acceptance model

The Technology Acceptance Model (TAM) by Davis (54) represents a central approach in acceptance research and serves as a foundation for many subsequent models (55). This theory aims to explain the individual usage behavior of new technologies and their influencing factors, with the goal of predicting technology use. This model is based on two significant influencing factors: ‘perceived usefulness’ and ‘perceived ease of use’ of the technology. Perceived usefulness refers to the extent to which individuals believe that using a particular system enhances their performance (54). Therefore, a high level of perceived usefulness is believed to positively influence users’ performance. Perceived ease of use describes the extent to which individuals assume that using a system requires little effort (54). Individuals evaluate a technology with a high perceived ease of use as simple to operate (54). Each scale consisted of six questions. The two subscales Perceived Usefulness and Perceived Ease of Use demonstrated excellent internal consistency, with Cronbach’s alphas of.98 and.94, respectively (54). Additionally, two further questions were included to assess the intention to use the app. The questionnaire consisted of 14 items, which were answered using a seven-point Likert scale (1 = “Strongly Disagree” to 7 = “Strongly agree”).

2.9 Analysis

All statistical analyses were performed using IBM SPSS Statistics software (version 30) (56). Demographic characteristics were analyzed using descriptive statistics, including means (M) and standard deviations (SD). The usability and acceptance scales were evaluated exclusively at the second measurement point (t₁). To examine group and sex differences in the questionnaire outcomes, different statistical approaches were applied based on the structure of the respective instruments. For questionnaires consisting of a single overall scale (i.e., SUS and the overall score of the TUI), a one-way analysis of co-variance (ANCOVA) was performed. For instruments comprising multiple subscales (i.e., TAM and the subscales of the TUI), a multivariate analysis of co-variance (MANCOVA) was conducted. Statistical significance was set at p <.05 (two-tailed hypothesis testing) for all analyses. Prior to conducting ANCOVA and MANCOVA, all statistical assumptions were tested and met. To account for potential baseline imbalances between group and sex, ANCOVAs and MANCOVAs were adjusted for baseline depression severity (BDI-II t₀ and MADRS t₀ scores). Adjusted effect sizes (partial η²) and 95% confidence intervals (CIs) are reported. Bonferroni corrections were applied to control for multiple comparisons in both the ANCOVA and MANCOVA analyses.

3 Results

3.1 Characteristics of the participants

The study included 32 inpatients with moderate to high depression symptom severity according to ICD-10 criteria. The mean age of participants was 38.12 years (SD = 14.03; range = 19–65), with an equal distribution of women (n = 16) and men (n = 16) across the SAR group (with motivational feedback) and tablet group (without motivational feedback). Overall, the participants presented with moderate depressive symptomatology, as indicated by mean MADRS scores of 30.63 (± 4.99) and BDI-II scores of 25.44 (± 8.57). The groups were comparable in age and sex composition, while the SAR group (with motivational feedback) showed slightly higher baseline depression severity than the tablet group (without motivational feedback) (see Table 1). No participants dropped out during the study period, and all completed both training sessions and questionnaires. Data were checked for completeness before analysis; no cases with missing values were identified. Therefore, all 32 participants were included in the final analyses. Mean scores (M), standard deviations and (SD) and 95% confidence intervals of the SUS, the TUI and the TAM are presented in Table 2.

Table 1

Table 1. Demographic and clinical characteristics of the sample.

Table 2

Table 2. Descriptive statistics of system usability scale, technology usage inventory and technology acceptance modell between SAR and tablet group and sex.

3.2 System usability scale

Bonferroni-corrected post-hoc analysis revealed a significant difference between SUS scores of the SAR group (with motivational feedback) and the tablet group (without motivational feedback (p <.001, M_Diff = 15.317, 95%-CI[6.938, 23.696]), after adjusting for baseline depression scores (BDI-II t₀ and MADRS t₀). No statistically significant results were found for sex comparisons (see Table 3).

Table 3

Table 3. Comparison of system usability scale and technology usage inventory overall between SAR and tablet group and sex after adjusting for baseline depression scores (BDI-II t₀ and MADRS t₀).

3.3 Technology usage inventory

Bonferroni-corrected post-hoc analysis revealed a significant difference between TUI overall scores of females and males (p = .033, M_Diff = 11.479, 95%-CI [.974, 21.983]), after adjusting for baseline Depression scores (BDI-II t₀ and MADRS t₀) (see Table 3).

After adjusting for baseline depression scores (BDI-II t₀ and MADRS t₀), a multivariate analysis of covariance (MANCOVA) revealed significant group differences for two subscales of the TUI.

Participants in the SAR group (with motivational feedback) reported significantly higher scores in the User-Friendliness subscale compared to the Tablet group (without motivational feedback), F (1, 28) = 12.18, p = .002, η²_p = .303, 95% CI [1.81, 6.95]. In contrast, the tablet group showed significantly higher scores in the subscale Accessibility than the SAR group, F (1, 28) = 8.07, p = .008, η²_p = .224, 95% CI [–6.70, –1.09].

A further significant effect was found for sex in the subscale Scepticism, with female participants reporting higher scores than male participants, F(1, 28) = 4.34, p = .046, η²_p = .134, 95% CI [0.06, 6.90].

3.4 Technology acceptance model

After adjusting for baseline depression scores (BDI-II t₀ and MADRS t₀), a multivariate analysis of covariance (MANCOVA) revealed no statistically significant results for all subscale for group and sex comparisons (see Table 4).

Table 4

Table 4. Comparison of the technology usage inventory and the technology acceptance model between SAR and tablet group and sex after adjusting for baseline depression scores (BDI-II t₀ and MADRS t₀).

4 Discussion

This study aimed to evaluate the usability and acceptance of the SAR “Pepper” when implemented as a motivational adjunct to tablet-based cognitive training, compared to cognitive training delivered without robotic support, in inpatients with depression, with a particular focus on sex-specific differences. These findings provide valuable insights into the integration of robotic technologies into psychiatric care. Given the increasing interest in technology-based interventions for mental health care, understanding user preferences and technological acceptance is crucial for the successful implementation of such interventions (8).

Regarding technology usability and acceptance, the results revealed several significant findings. Usability ratings measured via the SUS and user-friendliness measured via TUI User-Friendliness favored the SAR group (with motivational feedback), with significantly higher scores compared to the tablet group (without motivational feedback). In contrast the tablet group (with motivational feedback) showed significantly higher scores in the TUI subscale Accessibility. Additionally, sex-related differences were observed, with female participants demonstrating significantly higher TUI Overall and Scepticism subscale scores than males.

These findings highlight that while the robot-assisted intervention was perceived as usable and user-friendly, it was also met with higher skepticism. This finding indicates that higher usability does not necessarily translate into greater overall acceptance. The SAR condition appears to evoke ambivalent reactions, as participants appreciated the interactive and social features while simultaneously expressing reservations regarding the SAR interaction.

SAR “Pepper’s” usability was also rated positively, indicating a user-friendly design for human-robot interaction for patients with depression. These results align with previous research, indicating that SARs tend to be rated as highly engaging due to their interactive nature and human-like characteristics (8, 57). For example, the robot ‘Ryan’, which delivers internet-based cognitive behavioral therapy, has been shown to improve mood scores and engagement in older adults with depression (9). This underscores the importance of intuitive usability for the successful integration of robotic technology into clinical environments. Similar findings were reported by (11), who demonstrated that high usability improved SAR acceptance and utilization among older adults. Notably, patients with depression, who often face cognitive impairments (16), reported no significant difficulties in operating the robot. The user-friendly approach of SAR “Pepper” may be attributed to its perception as a supportive complement to traditional therapy rather than a replacement for human interaction (5). This interpretation is supported by Göransson et al. (58), who found that patients evaluated SARs more positively when they were presented as an addition rather than a substitute for human contact. Broadbent (3) emphasized the importance of intuitive and user-friendly design in overcoming barriers to technology use and facilitating its integration into clinical settings.

Qualitative feedback collected informally after the sessions supported these findings: several participants in the SAR group described SAR “Pepper” as “interesting but somewhat unusual at first,” and some expressed uncertainty about “how natural” the interaction felt compared with human contact. Others reported technical aspects such as delayed speech recognition or limited gesture responsiveness as minor obstacles to immersion. Similar concerns have been described in previous research, where users reported feelings of unfamiliarity, artificiality, or reduced authenticity in early stages of human–robot interaction (59, 60).

Furthermore, these findings align with those of Bishop et al. (17), who documented a generally positive attitude towards healthcare robots. Da Rocha Ferreira et al. (5) also highlighted the importance of clear communication regarding the added value of SAR in driving their acceptance. The results also showed that interacting with “Pepper” was perceived as enjoyable and motivating. This is particularly important because motivation and engagement play central roles in treating depression. The positive perception of usability underscores the importance of well-thought-out human-robot interactions to foster the acceptance and practical use of SAR.

Despite these advantages, skepticism toward SARs remains a notable barrier to acceptance. In our study, females reported significantly higher skepticism scores than men. This finding aligns with research suggesting that while SARs may be initially engaging, long-term usability and trust issues may arise, particularly among individuals with higher levels of anxiety or resistance to new technologies (61). Similar concerns were raised in studies evaluating the robot ‘Paro’, where some participants expressed reluctance to engage with a robotic intervention, perceiving it as artificial or intrusive (10, 62).

The significantly higher accessibility scores reported by participants in the tablet group (without motivational feedback) compared to those in the SAR group (with motivational feedback) can be attributed to a combination of cost- and usability-related factors. Tablets are generally more affordable and readily available, making them more accessible for a broader user base (63, 64). In contrast, SARs involve high upfront costs because of their complex design, restricting their use primarily to high-income settings (65). From a usability perspective, tablets benefit from familiar touchscreen interfaces and intuitive navigation, although certain design aspects, such as small screen size or complex applications, can potentially hinder their effectiveness for users with physical or cognitive limitations (66). Although SARs are specifically designed for ease of use, their technical limitations, including the need for structured training, may negatively impact perceived accessibility among older or mentally ill users (67). Overall, these findings underscore that the tablet’s combination of economic feasibility and relative usability likely contributed to its higher accessibility ratings in the current study.

A central aspect of this study was the exploration of sex-specific differences in the usability and acceptance of SAR “Pepper”. The results revealed that women tended to have a higher overall usability of SARs than men. This observation is consistent with the findings of Kuo et al. (26), who examined age- and sex-specific differences in the perception and acceptance of healthcare robots. Moradbekhti et al. (27) also found that women rated SARs more positively, particularly regarding its usefulness and user-friendliness. The reasons for these sex-specific differences are complex. Ringwald et al. (28) argue that such differences may be attributed to distinct socialization experiences and role expectations. Women, often characterized by stronger caregiving orientations and social sensitivity, may be more receptive to using supportive technologies. Additionally, they may place greater emphasis on the emotional and social aspects of interacting with SARs than men. This suggests that the relationship between sex and SAR acceptance is complex and potentially context-dependent. The sex-specific findings of this study highlight the necessity of incorporating sex-sensitive approaches into the development and implementation of SARs. This could involve targeted adjustments in human-robot interaction or differentiated communication strategies that address the specific needs and expectations of different sexes.

Another critical aspect of SARs adoption is the role of sex-specific preferences. Recent research suggests that men and women may have different expectations of SARs, with women placing greater emphasis on emotional support and social interaction and men prioritizing functional aspects and usability (68). This may explain some of the variability in user engagement and skepticism observed in the present study. Furthermore, studies have highlighted that the personality adaptation of SARs can significantly influence user acceptance. For example, SARs with extroverted personalities tend to be more accepted by extroverted users, whereas introverted users may prefer calmer, less expressive robotic interactions (69, 70). These findings emphasize the importance of adaptive, user-tailored robot interactions, a concept that has been incorporated into SAR interventions, such as this project (35).

A recurring challenge in SAR-based interventions is the sustainability of engagement over time. While our study found a preference for the SAR over the tablet in terms of usability and user friendliness, it remains unclear whether this preference would persist over an extended period. Previous research indicates that SAR engagement often decreases after an initial novelty phase, leading to reduced long-term adherence (10, 62). A possible explanation for this phenomenon is the limited adaptability of SARs, which may lead users to perceive interactions as repetitive or lacking meaningful engagement over time (12).

Additionally, ethical concerns regarding privacy, data security, and autonomy have been raised in previous studies, particularly in relation to psychiatric patients (61). While SARs offer promising solutions for cognitive training and emotional support, their successful implementation in clinical practice requires addressing the above-mentioned barriers to trust and acceptance.

The positive results regarding the usability and acceptance of SAR “Pepper” among patients with depression highlight the promising opportunities for SARs use in psychiatric care. SARs can serve as a complementary tool in depression treatment, assisting with cognitive training exercises, providing information, and offering emotional support. Studies such as those by Chen et al. (71) have demonstrated that SAR-based intervention programs can positively affect cognitive function and mood in older adults in long-term care facilities. Similarly, Chen et al. (72) reported successful SAR interventions in children with autism spectrum disorders. These findings suggest that SARs could be versatile tools for treating depression.

5 Strengths and limitations

This study provides valuable insights into the usability and acceptance of SARs and tablet-based cognitive training in individuals with depression. A key strength of this study is the randomized group assignment, which minimizes selection bias and ensures comparability between the intervention groups. By using validated psychometric instruments, such as the MADRS and BDI-II, this study offers robust clinical assessments of depressive symptoms, increasing the reliability of the findings.

Another strength of this study is the use of multiple standardized measures to assess technology usability and acceptance. The inclusion of the SUS, the TUI, and the TAM allows for a comprehensive evaluation of the user experience.

Additionally, this study contributes to the growing body of research on SARs in mental health care by examining their clinical usability. Previous research has mainly focused on older adults or individuals with cognitive impairments (9, 57), whereas this study specifically addressed depression in an adult population, thereby expanding the applicability of SAR interventions to broader psychiatric settings.

Despite its promising findings and strengths, this study has several limitations that should be acknowledged. The main limitation concerns the asymmetry in feedback between the two groups. The SAR group included verbal motivational feedback of the SAR “Pepper”, whereas the tablet condition did not provide any motivational feedback beyond task progression. This introduces an additional factor, that may have contributed to the higher usability and acceptance ratings in the SAR group. Consequently, the results should be interpreted as reflecting the combined effect of humanoid interaction and motivational feedback rather than the robotic embodiment alone. Future studies should equalize feedback across conditions to isolate the specific contribution of the SAR component.

The second limitations was the relatively small the sample size (n = 32), which may limit the generalizability of the findings. While the randomized design strengthens internal validity, the statistical power of the study remains restricted, particularly in detecting subtle effects or interaction effects between the variables. Future studies should aim to replicate these findings in larger and more diverse samples to enhance their external validity.

Another limitation relates to the limited intervention duration and training frequency of the cognitive training intervention. Participants completed only two training sessions, which may not have been sufficient to produce measurable improvements in cognitive performance. Prior studies have indicated that long-term engagement is crucial for the sustained benefits of SAR-based interventions, as engagement levels may decline after an initial novelty phase (10, 62). Abdi et al. (73) emphasized the importance of long-term studies to better understand the lasting impact and acceptance of SARs in healthcare settings. Future research should explore the effects of prolonged SAR interventions on cognitive and emotional outcomes.

The sex-specific differences observed in this study warrant further investigation in future studies. Intersectional aspects should also be considered to develop a nuanced understanding of the factors influencing SAR usability and its acceptance. Eyssel and Hegel (74) emphasized the need to incorporate cultural and socioeconomic factors into the analysis to obtain a holistic perspective on this issue.

This study focused exclusively on patients with depression. Future research should explore the use of SARs for other psychiatric conditions, such as anxiety disorders, bipolar disorders, and post-traumatic stress disorders. Studies by Huijnen et al. (22) and Rabbitt et al. (75) have shown promising results for the use of SARs among patients with autism spectrum disorders and other psychiatric diseases.

Beyond the methodological and conceptual limitations discussed above, two additional aspects warrant further attention. The first relates to the integration of SARs into existing treatment frameworks. While this study focused on the evaluation of usability and acceptance, it did not examine how SAR-based cognitive training interacts with standard therapeutic approaches. Therefore, future studies should investigate whether combining SAR interventions with conventional treatments can enhance therapeutic outcomes. As highlighted by Tapus et al. (76), such integrative approaches are crucial for harnessing the full potential of SARs in psychiatric care.

Another important consideration is the ethical dimensions of SAR use in sensitive clinical environments. Although not the focus of this study, issues such as data privacy, informed consent, and unintended psychosocial consequences of human-robot interactions are highly relevant, particularly when working with vulnerable populations. Addressing these questions in future research is essential to ensure ethical implementation. Sharkey and Sharkey (77) emphasize the importance of proactively engaging with ethical challenges in the context of healthcare robotics.

6 Conclusion

The overall positive evaluation of SAR “Pepper” underscores the potential of SARs as a complementary tool in psychiatric care. The observed sex-specific differences highlight the need for tailored approaches to developing and implementing SAR technology in clinical environments.

The results of this study provide a solid foundation for further research and advancement of SAR technology in psychiatric care. They highlight the potential of SAR technology to enhance treatment quality and open new pathways for patient care. Simultaneously, they caution against an overreliance on SAR and stress the need to view them as complements, not replacements, to human interaction and traditional therapies.

Future research should investigate the long-term effects of SAR in psychiatric care, extend its applicability to other mental health conditions, and explore complex interactions between SAR technology and traditional treatment approaches. Only through continuous research and careful evaluation can we fully harness the potential of SAR in psychiatric care while ensuring their use remains ethical and patient centered.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

Ethics statement

The studies involving humans were approved by Ethics Committee of the Medical University of Graz (EK Nr. 35-450 ex 22/23). The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation in this study was provided by the participants’ legal guardians/next of kin. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

Author contributions

AH: Conceptualization, Data curation, Formal Analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Validation, Visualization, Writing – original draft, Writing – review & editing. IZ: Data curation, Investigation, Resources, Writing – review & editing. LS: Data curation, Resources, Writing – review & editing. IS: Data curation, Resources, Writing – review & editing. MSt: Data curation, Resources, Writing – review & editing. TS: Formal Analysis, Writing – review & editing. ML: Conceptualization, Funding acquisition, Project administration, Writing – review & editing. FF: Conceptualization, Funding acquisition, Project administration, Writing – review & editing. SG: Funding acquisition, Project administration, Writing – review & editing. ES: Resources, Writing – review & editing. MP: Conceptualization, Funding acquisition, Writing – review & editing, Data curation, Methodology. TO: Conceptualization, Funding acquisition, Writing – review & editing, Data curation, Methodology. SD: Conceptualization, Funding acquisition, Writing – review & editing, Methodology. MSc: Conceptualization, Funding acquisition, Writing – review & editing, Methodology. SR: Conceptualization, Funding acquisition, Project administration, Writing – review & editing, Methodology, Supervision. JZ: Conceptualization, Funding acquisition, Methodology, Project administration, Supervision, Writing – review & editing. DS: Data curation, Writing – review & editing. AS: Data curation, Writing – review & editing. SS: Conceptualization, Funding acquisition, Methodology, Project administration, Supervision, Writing – review & editing. MM: Project administration, Writing – review & editing, Conceptualization, Funding acquisition, Methodology, Supervision. LP: Conceptualization, Funding acquisition, Project administration, Writing – review & editing, Resources, Supervision, Methodology. ND: Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Writing – review & editing, Methodology. ER: Conceptualization, Funding acquisition, Methodology, Project administration, Resources, Supervision, Writing – review & editing.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This work was partly funded by project AMIGA under grant No FO999892829 of the Austrian FFG femtech programme co-financed by the Austrian Federal Ministry for Climate Action, Environment, Energy, Mobility, Innovation and Technology (BMK). #AH will receive funding to cover the publication fee costs for this article from the Doctoral School “Lifestyle-related Diseases” at the Medical University of Graz.

Acknowledgments

The authors thank all the participants for their cooperation in this study. Additionally, we would like to express our gratitude to Humanizing Technologies GmbH, especially Dimitrios Prodromou, for their support in the technical implementation of the SAR “Pepper.” This publication was prepared in partial fulfilment of the requirements for the doctoral program in Medical Science (doctoral school “Lifestyle Related Diseases”) at the Medical University of Graz, Austria.

Conflict of interest

Author MM was employed by Humanizing Technologies GmbH.

The remaining author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The authors declare that Gen AI was used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

1. World Health Organization (WHO). Monitoring mental health systems and services in the WHO European Region: Mental Health Atlas. World Health Organization. Regional Office for Europe. (2019). Available online at: https://iris.who.int/handle/10665/346404.

Google Scholar

2. Malhi GS, Bell E, Bassett D, Boyce P, Bryant R, Hazell P, et al. The 2020 Royal Australian and New Zealand College of Psychiatrists clinical practice guidelines for mood disorders. Aust N Z J Psychiatry. (2021) 55:7–117. doi: 10.1177/0004867420979353

PubMed Abstract | Crossref Full Text | Google Scholar

3. Broadbent E. Interactions with robots: the truths we reveal about ourselves. Annu Rev Psychol. (2017) 68:627–52. doi: 10.1146/annurev-psych-010416-043958

PubMed Abstract | Crossref Full Text | Google Scholar

4. Scoglio AA, Reilly ED, Gorman JA, and Drebing CE. Use of social robots in mental health and well-being research: systematic review. J Med Internet Res. (2019) 21:e13322. doi: 10.2196/13322

PubMed Abstract | Crossref Full Text | Google Scholar

5. Da Rocha Ferreira AR, Silva MGMMDS, De Oliveira Correia M, De Almeida RS, Simões-Silva V, and Trigueiro MJ. Use of Socially Assistive Robots in Mental Health: Barriers and Facilitators. In: Simões De Almeida R, Simões-Silva V, and Trigueiro MJ, editors. Handbook of Research on Advances in Digital Technologies to Promote Rehabilitation and Community Participation. IGI Global Scientific Publishing (2023). p. 269–92. doi: 10.4018/978-1-6684-9251-2.ch013

Crossref Full Text | Google Scholar

6. Spitale M, Okamoto S, Gupta M, Xi H, and Matarić MJ. Socially assistive robots as storytellers that elicit empathy. ACM Trans Hum-Robot Interact. (2022) 11:1–29. doi: 10.1145/3538409

Crossref Full Text | Google Scholar

7. Yee LOH, Fun SS, Zin TS, Aung ZN, Yap KM, and Teoh J. Socially assistive robots: A technological approach to emotional support. (2024). doi: 10.48550/ARXIV.2411.05122

Crossref Full Text | Google Scholar

8. Lee HR, Šabanović S, Chang W-L, Nagata S, Piatt J, Bennett C, et al. (2017). Steps toward participatory design of social robots: mutual learning with older adults with depression, in: Proceedings of the 2017 ACM/IEEE International Conference on Human-Robot Interaction, Vienna Austria. pp. 244–53. doi: 10.1145/2909824.3020237

Crossref Full Text | Google Scholar

9. Dino F, Zandie R, Abdollahi H, Schoeder S, and Mahoor MH. (2019). Delivering cognitive behavioral therapy using A conversational social robot, in: 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Macau, China. pp. 2089–95. doi: 10.1109/IROS40897.2019.8968576

Crossref Full Text | Google Scholar

10. Randall N, Bennett CC, Šabanović S, Nagata S, Eldridge L, Collins S, et al. More than just friends: in-home use and design recommendations for sensing socially assistive robots (SARs) by older adults with depression. Paladyn J Behav Robot. (2019) 10:237–55. doi: 10.1515/pjbr-2019-0020

Crossref Full Text | Google Scholar

11. Khosla R, Nguyen K, and Chu M-T. Human robot engagement and acceptability in residential aged care. Int J Human–Computer Interact. (2017) 33:510–22. doi: 10.1080/10447318.2016.1275435

Crossref Full Text | Google Scholar

12. Guemghar I, Pires de Oliveira Padilha P, Abdel-Baki A, Jutras-Aswad D, Paquette J, and Pomey M-P. Social robot interventions in mental health care and their outcomes, barriers, and facilitators: scoping review. JMIR Ment Health. (2022) 9:e36094. doi: 10.2196/36094

PubMed Abstract | Crossref Full Text | Google Scholar

13. Rantanen T, Lehto P, Vuorinen P, and Coco K. The adoption of care robots in home care—A survey on the attitudes of Finnish home care personnel. J Clin Nurs. (2018) 27:1846–59. doi: 10.1111/jocn.14355

PubMed Abstract | Crossref Full Text | Google Scholar

14. Papadopoulos I, Wright S, Koulouglioti C, Ali S, Lazzarino R, Martín-García Á, et al. Socially assistive robots in health and social care: Acceptance and cultural factors. Results from an exploratory international online survey. Jpn J Nurs Sci. (2023) 20:e12523. doi: 10.1111/jjns.12523

PubMed Abstract | Crossref Full Text | Google Scholar

15. Moharana S, Panduro AE, Lee HR, and Riek LD. (2019). Robots for joy, robots for sorrow: community based robot design for dementia caregivers, in: 2019 14th ACM/IEEE International Conference on Human-Robot Interaction (HRI), Daegu, Korea (South. pp. 458–67. doi: 10.1109/HRI.2019.8673206

Crossref Full Text | Google Scholar

16. Ryan M and Stahl BC. Artificial intelligence ethics guidelines for developers and users: clarifying their content and normative implications. J Inf Commun Ethics Soc. (2021) 19:61–86. doi: 10.1108/JICES-12-2019-0138

Crossref Full Text | Google Scholar

17. Bishop L, Van Maris A, Dogramadzi S, and Zook N. Social robots: The influence of human and robot characteristics on acceptance. Paladyn J Behav Robot. (2019) 10:346–58. doi: 10.1515/pjbr-2019-0028

Crossref Full Text | Google Scholar

18. Paletta L, Brijacak I, Reiterer B, Pszeida M, Ganster H, Fuhrmann F, et al. (2019). Gaze-based human factors measurements for the evaluation of intuitive human-robot collaboration in real-time, in: 2019 24th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA), Zaragoza, Spain. pp. 1528–31. doi: 10.1109/ETFA.2019.8869270

Crossref Full Text | Google Scholar

19. Paletta L, Schüssler S, Zuschnegg J, Steiner J, Pansy-Resch S, Lammer L, et al. AMIGO—A socially Assistive Robot for Coaching Multimodal Training of Persons with Dementia. In: Korn O, editor. Social Robots: Technological, Societal and Ethical Aspects of Human-Robot Interaction. Springer International Publishing, Cham (2019). p. 265–84. doi: 10.1007/978-3-030-17107-0_13

Crossref Full Text | Google Scholar

20. Zuschnegg J, Schoberer D, Häussl A, Herzog SA, Russegger S, Ploder K, et al. Effectiveness of computer-based interventions for community-dwelling people with cognitive decline: a systematic review with meta-analyses. BMC Geriatr. (2023) 23:229. doi: 10.1186/s12877-023-03941-y

PubMed Abstract | Crossref Full Text | Google Scholar

21. Pu L, Moyle W, Jones C, and Todorovic M. The effectiveness of social robots for older adults: A systematic review and meta-analysis of randomized controlled studies. Gerontologist. (2019) 59:e37–51. doi: 10.1093/geront/gny046

PubMed Abstract | Crossref Full Text | Google Scholar

22. Huijnen CA, As Lexis M, and De Witte LP. Robots as new tools in therapy and education for children with autism. Int J Neurorehabilitation. (2017) 04. doi: 10.4172/2376-0281.1000278

Crossref Full Text | Google Scholar

23. Torous J, Bucci S, Bell IH, Kessing LV, Faurholt-Jepsen M, Whelan P, et al. The growing field of digital psychiatry: current evidence and the future of apps, social media, chatbots, and virtual reality. World Psychiatry. (2021) 20:318–35. doi: 10.1002/wps.20883

PubMed Abstract | Crossref Full Text | Google Scholar

24. Lipschitz JM, Pike CK, Hogan TP, Murphy SA, and Burdick KE. The engagement problem: a review of engagement with digital mental health interventions and recommendations for a path forward. Curr Treat Options Psychiatry. (2023) 10:119–35. doi: 10.1007/s40501-023-00297-3

PubMed Abstract | Crossref Full Text | Google Scholar

25. Guerrera CS, Platania GA, Boccaccio FM, Sarti P, Varrasi S, Colliva C, et al. The dynamic interaction between symptoms and pharmacological treatment in patients with major depressive disorder: the role of network intervention analysis. BMC Psychiatry. (2023) 23:885. doi: 10.1186/s12888-023-05300-y

PubMed Abstract | Crossref Full Text | Google Scholar

26. Kuo IH, Rabindran JM, Broadbent E, Lee YI, Kerse N, Stafford RMQ, et al. (2009). Age and gender factors in user acceptance of healthcare robots, in: RO-MAN 2009 - The 18th IEEE International Symposium on Robot and Human Interactive Communication, Toyama, Japan. pp. 214–9. doi: 10.1109/ROMAN.2009.5326292

Crossref Full Text | Google Scholar

27. Moradbakhti L, Mara M, Castellano G, and Winkle K. (Counter-)stereotypical gendering of robots in care: impact on needs satisfaction and gender role concepts in men and women users. Int J Soc Robot. (2023) 15:1769–90. doi: 10.1007/s12369-023-01033-w

Crossref Full Text | Google Scholar

28. Ringwald M, Theben P, Gerlinger K, Hedrich A, and Klein B. How should your assistive robot look like? A scoping review on embodiment for assistive robots. J Intell Robot Syst. (2023) 107:12. doi: 10.1007/s10846-022-01781-3

Crossref Full Text | Google Scholar

29. Hopewell S, Chan A-W, Collins GS, Hróbjartsson A, Moher D, Schulz KF, et al. CONSORT 2025 statement: updated guideline for reporting randomised trials. BMJ. (2025), e081123. doi: 10.1136/bmj-2024-081123

PubMed Abstract | Crossref Full Text | Google Scholar

30. Hengartner MP, Jakobsen JC, Sørensen A, and Plöderl M. Efficacy of new-generation antidepressants assessed with the Montgomery-Asberg Depression Rating Scale, the gold standard clinician rating scale: A meta-analysis of randomised placebo-controlled trials. PloS One. (2020) 15:e0229381. doi: 10.1371/journal.pone.0229381

PubMed Abstract | Crossref Full Text | Google Scholar

31. Strauß B and Schumacher J. Klinische interviews und ratingskalen. Göttingen: Hogrefe Verlag (2005).

Google Scholar

32. Hautzinger M, Keller F, and Kühner C. BDI-II. Beck-depressions-inventar. Frankfurt: Pearson Assessment (2009).

Google Scholar

33. Maher JM, Markey JC, and Ebert-May D. The other half of the story: effect size analysis in quantitative research. CBE Life Sci Educ. (2013) 12:345–51. doi: 10.1187/cbe.13-04-0082

PubMed Abstract | Crossref Full Text | Google Scholar

34. Pszeida M, Häussl A, Zuschnegg J, Macher M, Draxler S, Orgel T, et al. Motivating patients with depression for gender-sensitive cognitive training using a socially assistive robot with bio-signal driven pause management. (2025) 165. doi: 10.54941/ahfe1006090

Crossref Full Text | Google Scholar

35. Paletta L, Schuessler S, Reininghaus E, Macher M, Fellner M, Russegger S, et al. Towards gender-sensitive motivation of patients with depression for cognitive training with a socially assistive robot. (2023) 73. doi: 10.54941/ahfe1003969

Crossref Full Text | Google Scholar

36. Faul F, Erdfelder E, Buchner A, and Lang A-G. Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses. Behav Res Methods. (2009) 41:1149–60. doi: 10.3758/BRM.41.4.1149

PubMed Abstract | Crossref Full Text | Google Scholar

37. Faul F, Erdfelder E, Lang A-G, and Buchner A. G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods. (2007) 39:175–91. doi: 10.3758/BF03193146

PubMed Abstract | Crossref Full Text | Google Scholar

38. Study randomizer. Available online at: https://www.studyrandomizer.com (Accessed July 4, 2025).

Google Scholar

39. LimeSurvey GmbH. LimeSurvey: an open source survey tool (2025). Available online at: https://www.limesurvey.org (Accessed July 4, 2025).

Google Scholar

40. Aldebaran. Aldebaran - pepper. Marseille, France: Aldebaran (2025). Available online at: https://aldebaran.com/en/pepper/ (Accessed June 6, 2025).

Google Scholar

41. Probo Robotics. Technische spezifikationen (2022). Available online at: https://www.probo-robotics.at/pepper-daten-fakten/ (Accessed November 22, 2022).

Google Scholar

42. Softbank Robotics. Pepper – The worlds first Personal Robot that Reads Emotions (2021). Available online at: https://www.softbankrobotics.com/emea/de/pepper (Accessed July 4, 2025).

Google Scholar

43. Feil-Seifer D and Matarić MJ. (2005). Defining socially assistive robotics, in: 9th Int Conf Rehabil Robot 2005 ICORR, . pp. 465–8.

Google Scholar

44. Paletta L, Fellner M, Pszeida M, Lerch A, Kemp C, Pittino L, et al. (2018). Playful multimodal training for persons with dementia with executive function based decision support, in: Proceedings of the 11th PErvasive Technologies Related to Assistive Environments Conference, Corfu Greece. pp. 237–40. doi: 10.1145/3197768.3197789

Crossref Full Text | Google Scholar

45. Paletta L, Pszeida M, Fellner M, Russegger S, Dini A, Draxler S, et al. Towards decision support with assessment of neuropsychological profiles in alzheimer’s dementia using playful tablet-based multimodal activation. In: Ayaz H, Asgher U, and Paletta L, editors. Advances in neuroergonomics and cognitive engineering. Springer International Publishing, Cham (2021). p. 469–77. doi: 10.1007/978-3-030-80285-1_54

Crossref Full Text | Google Scholar

46. dZuschnegg J, Schuessler S, Paletta L, Russegger S, Fellner M, Ploder K, et al. Usability evaluation after a 6-month tablet-based dementia training program by people with alzheimer’s disease, relatives, and dementia trainers. (2022) 43. doi: 10.54941/ahfe1001844

Crossref Full Text | Google Scholar

47. Fluet G, Qiu Q, Gross A, Gorin H, Patel J, Merians A, et al. The influence of scaffolding on intrinsic motivation and autonomous adherence to a game-based, sparsely supervised home rehabilitation program for people with upper extremity hemiparesis due to stroke. A randomized controlled trial. J NeuroEngineering Rehabil. (2024) 21:143. doi: 10.1186/s12984-024-01441-7

PubMed Abstract | Crossref Full Text | Google Scholar

48. Bangor A, Kortum PT, and Miller JT. An empirical evaluation of the system usability scale. Int J Hum-Comput Interact. (2008) 24:574–94. doi: 10.1080/10447310802205776

Crossref Full Text | Google Scholar

49. Brooke J. SUS: A quick and dirty usability scale. Usability Eval Ind. (1995) 189.

Google Scholar

50. Lewis JR. The system usability scale: past, present, and future. Int J Human–Computer Interact. (2018) 34:577–90. doi: 10.1080/10447318.2018.1455307

Crossref Full Text | Google Scholar

51. Khan Q, Hickie IB, Loblay V, Ekambareshwar M, Zahed IUM, Naderbagi A, et al. Psychometric evaluation of the System Usability Scale in the context of a childrearing app co-designed for low- and middle-income countries. Digit Health. (2025) 11:20552076251335413. doi: 10.1177/20552076251335413

PubMed Abstract | Crossref Full Text | Google Scholar

52. Sauro J. A practical guide to the system usability scale: background, benchmarks & best practices. Denver, CO: Measuring Usability LLC (2011). p. 162.

Google Scholar

53. Kothgasser O, Felnhofer A, Hauk N, Kastenhofer E, Gomm J, and Kryspin-Exner I. TUI - technology usage inventory manual (2012). Available online at: https://www.ffg.at/sites/default/files/allgemeine_downloads/thematische%20programme/programmdokumente/tui_manual.pdf (Accessed July 4, 2025).

Google Scholar

54. Davis FD. Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Q. (1989) 13:319. doi: 10.2307/249008

Crossref Full Text | Google Scholar

55. Venkatesh V and Davis FD. A theoretical extension of the technology acceptance model: four longitudinal field studies. Manag Sci. (2000) 46:186–204. doi: 10.1287/mnsc.46.2.186.11926

Crossref Full Text | Google Scholar

56. IBM Corp. IBM SPSS statistics for windows. Armonk, NY: IBM Corp. (2023).

Google Scholar

57. Šabanović S, Chang W-L, Bennett CC, Piatt JA, and Hakken D. A robot of my own: participatory design of socially assistive robots for independently living older adults diagnosed with depression. In: Zhou J and Salvendy G, editors. Human aspects of IT for the aged population. Design for aging. Springer International Publishing, Cham (2015). p. 104–14. doi: 10.1007/978-3-319-20892-3_11

Crossref Full Text | Google Scholar

58. Göransson O, Pettersson K, Larsson PA, and Lennernäs B. Personals attitudes towards robot assisted health care – a pilot study in 111 respondents. Stud Health Technol Inform. (2008) 137:56–60.

PubMed Abstract | Google Scholar

59. Laban G, Kappas A, Morrison V, and Cross ES. Building long-term human–robot relationships: examining disclosure, perception and well-being across time. Int J Soc Robot. (2024) 16:1–27. doi: 10.1007/s12369-023-01076-z

Crossref Full Text | Google Scholar

60. Rosén J, Lindblom J, Lamb M, and Billing E. Previous experience matters: an in-person investigation of expectations in human–robot interaction. Int J Soc Robot. (2024) 16:447–60. doi: 10.1007/s12369-024-01107-3

Crossref Full Text | Google Scholar

61. Mlakar I, Smrke U, Flis V, Kobilica N, Horvat S, Ilijevec B, et al. Using structural equation modeling to explore patients’ and healthcare professionals’ Expectations and attitudes towards socially assistive humanoid robots in nursing and care routine. Int J Soc Robot. (2024) 16:105–24. doi: 10.1007/s12369-023-01039-4

Crossref Full Text | Google Scholar

62. Piatt J, Nagata S, Šabanović S, Cheng W-L, Bennett C, Lee HR, et al. Companionship with a robot? Therapists’ perspectives on socially assistive robots as therapeutic interventions in community mental health for older adults. Am J Recreat Ther. (2017) 15:29. doi: 10.5055/ajrt.2016.0117

Crossref Full Text | Google Scholar

63. Becchimanzi C, Iacono E, and Brischetto A. Acceptability of assistive robotics by older adults: results from a human-centred qualitative study. (2022) 37. doi: 10.54941/ahfe1001637

Crossref Full Text | Google Scholar

64. Maranesi E, Amabili G, Cucchieri G, Bolognini S, Margaritini A, and Bevilacqua R. Understanding the acceptance of ioT and social assistive robotics for the healthcare sector: A review of the current user-centred applications for the older users. In: Scataglini S, Imbesi S, and Marques G, editors. Internet of things for human-centered design. Studies in computational intelligence. Springer Nature Singapore, Singapore (2022). p. 331–51. doi: 10.1007/978-981-16-8488-3_16

Crossref Full Text | Google Scholar

65. Motahar T, Farden MDF, Sarkar DP, Islam MDA, Cabrera ME, and Cakmak M. (2019). SHEBA: A low-cost assistive robot for older adults in the developing world, in: 2019 28th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN), New Delhi, India. pp. 1–8. doi: 10.1109/RO-MAN46459.2019.8956402

Crossref Full Text | Google Scholar

66. Hammer S, Kirchner K, André E, and Lugrin B. (2017). Touch or talk?: comparing social robots and tablet PCs for an elderly assistant recommender system, in: Proceedings of the Companion of the 2017 ACM/IEEE International Conference on Human-Robot Interaction, Vienna Austria. pp. 129–30.

Google Scholar

67. Mahmoudi Asl A, Toribio-Guzmán JM, Castro-González Á, Malfaz M, Salichs MÁ, van der Roest H, et al. Evaluating the user experience and usability of the MINI robot for elderly adults with mild dementia and mild cognitive impairment: insights and recommendations. (2024) 13:2081–2102. doi: 10.20944/preprints202408.1465.v1

PubMed Abstract | Crossref Full Text | Google Scholar

68. Forgas-Coll S, Huertas-Garcia R, Andriella A, and Alenyà G. Gendered human–robot interactions in services. Int J Soc Robot. (2023) 15:1791–807. doi: 10.1007/s12369-023-01035-8

Crossref Full Text | Google Scholar

69. Andriella A, Siqueira H, Fu D, Magg S, Barros P, Wermter S, et al. Do I have a personality? Endowing care robots with context-dependent personality traits. Int J Soc Robot. (2021) 13:2081–102. doi: 10.1007/s12369-020-00690-5

Crossref Full Text | Google Scholar

70. Ahmad R, Siemon D, Gnewuch U, and Robra-Bissantz S. Designing personality-adaptive conversational agents for mental health care. Inf Syst Front. (2022) 24:923–43. doi: 10.1007/s10796-022-10254-9

PubMed Abstract | Crossref Full Text | Google Scholar

71. Chen F-Q, Leng Y-F, Ge J-F, Wang D-W, Li C, Chen B, et al. Effectiveness of virtual reality in nursing education: meta-analysis. J Med Internet Res. (2020) 22:e18290. doi: 10.2196/18290

PubMed Abstract | Crossref Full Text | Google Scholar

72. Chen S, Jones C, and Moyle W. Social robots for depression in older adults: A systematic review. J Nurs Scholarsh. (2018) 50:612–22. doi: 10.1111/jnu.12423

PubMed Abstract | Crossref Full Text | Google Scholar

73. Abdi J, Al-Hindawi A, Ng T, and Vizcaychipi MP. Scoping review on the use of socially assistive robot technology in elderly care. BMJ Open. (2018) 8:e018815. doi: 10.1136/bmjopen-2017-018815

PubMed Abstract | Crossref Full Text | Google Scholar

74. Eyssel F and Hegel F. (S)he’s got the look: gender stereotyping of robots¹. J Appl Soc Psychol. (2012) 42:2213–30. doi: 10.1111/j.1559-1816.2012.00937.x

Crossref Full Text | Google Scholar

75. Rabbitt SM, Kazdin AE, and Scassellati B. Integrating socially assistive robotics into mental healthcare interventions: Applications and recommendations for expanded use. Clin Psychol Rev. (2015) 35:35–46. doi: 10.1016/j.cpr.2014.07.001

PubMed Abstract | Crossref Full Text | Google Scholar

76. Tapus A, Tapus C, and Mataric MJ. (2009). The use of socially assistive robots in the design of intelligent cognitive therapies for people with dementia, in: 2009 IEEE International Conference on Rehabilitation Robotics, Kyoto, Japan. pp. 924–9. doi: 10.1109/ICORR.2009.5209501

Crossref Full Text | Google Scholar

77. Sharkey A and Sharkey N. We need to talk about deception in social robotics! Ethics Inf Technol. (2021) 23:309–16. doi: 10.1007/s10676-020-09573-9

Crossref Full Text | Google Scholar