Abstract
This study assessed whether nine persons with advanced Alzheimer’s disease would learn to engage in leg responses (exercise) with the support of a technology-aided program, which provided (a) preferred stimulation contingent on the leg responses and (b) verbal reminders/prompts in case of no responding. The study was conducted according to a non-concurrent multiple baseline design across participants and involved sessions of 5 min. During the baseline, the participants’ mean frequencies of leg responses ranged from zero to slightly above two per session. During the intervention, those frequencies ranged from nearly 10 to nearly 17 per session. The mean frequencies of prompts varied across participants from about two to more than seven per session. In addition to the increase in leg responses, participants showed an increase in signs of positive personal involvement (e.g., smiles and positive verbalizations) during the intervention sessions as compared with the baseline sessions. The applicability and potential benefits of the program in daily contexts are discussed.
Introduction
Alzheimer’s disease is a neurodegenerative disorder that generally affects persons more than 65 years of age, making them increasingly less independent in their daily life and eventually depriving them of any functional skills (Ambrose, 2012; Bernick, Cummings, Raman, Sun, & Aisen, 2012; Melrose et al., 2011; Perri, Monaco, Fadda, Caltagirone, & Carlesimo, 2014; Sikkes et al., 2013; Soto et al., 2012; Spalletta et al., 2012; Wilson et al., 2012). Initially, they lose more complex abilities such as the appropriate use of finances, time, or communication means (e.g., with problems in dealing with payments and appointments or in making phone calls; Campbell et al., 2012; Cotrell, Wild, & Bader, 2006; Marson et al., 2000; Perilli et al., 2012). Subsequently, they also lose their abilities of orienting in familiar outdoor as well as indoor contexts and of performing common daily activities, such as preparing food items, or executing self-care routines (Caffò et al., 2014; Lancioni et al., 2012; Lancioni et al., 2014; Martyr & Clare, 2012). Eventually, they can become very passive and sedentary, with no ambulation and occupational engagement and virtually no communication exchanges (Buettner & Fitzsimmons, 2002; Colling, 2004; Lancioni, Singh, O’Reilly, Green, et al., 2013; McHugh, Gardstrom, Hiller, Brewer, & Diestelkamp, 2012).
Given the high prevalence, and apparently inescapable consequences of the disease, extensive efforts have been made to identify pharmacological and behavioral intervention approaches that could partially/temporarily contain its progression and possibly slow down the deterioration process (Ferrero-Arias et al., 2011; Ferris & Farlow, 2013; Hoffmann et al., 2013; Konrath, Passos Cdos, Klein, & Henriques, 2013; Perilli et al., 2013; Raglio et al., 2010; Rao, Chou, Bursley, Smulofsky, & Jezequel, 2014; Rive et al., 2012; Versijpt, 2014). The behavioral intervention approaches available include, among others, (a) reality orientation training and memory exercises to promote the persons’ cognitive and interactive functioning within their context (Boller, Jennings, Dieudonné, Verny, & Ergis, 2012; Cotelli, Manenti, Zanetti, & Miniussi, 2012; Zanetti et al., 2001), (b) technology-aided instructions to help the persons carry out multi-step (functional) daily activities and self-care routines (Lancioni et al., 2012; Lancioni et al., 2014; Perilli et al., 2013), (c) dance and psychomotor activation to improve the persons’ cognitive and social behavior (Guzmán-García, Mukaetova-Ladinska, & James, 2013; Hopman-Rock, 2000; Hopman-Rock, Staats, Tak, & Dröes, 1999), (d) music therapy to reduce the persons’ agitation and wandering (Fitzgerald-Cloutier, 1993; Janata, 2012), and (e) technology-aided stimulation and prompts to foster the persons’ engagement in mild physical exercise (e.g., arm raising; Lancioni et al., 2015).
The last intervention approach was developed for persons with advanced Alzheimer’s disease, who had lost their ambulation skills and were sedentary and largely inactive, and was to help them engage in exercise virtually independent of caregivers’ supervision (i.e., in a practically affordable manner; Staedtler & Nunez, 2015). The program was assessed with six participants and included preferred stimulation contingent on the target response (i.e., either arm raising or leg pedaling) and verbal reminders/prompts in case of no responding. Data showed that the program was effective with the participants displaying an increase in the target response and signs of positive personal involvement (e.g., smiles and verbalizations; cf. Christofoletti et al., 2011; Farina, Rusted, & Tabet, 2014; Fischer, Langner, Birbaumer, & Brocke, 2008; Hoffmann et al., 2013; Rolland, Abellan van Kan, & Vellas, 2008).
Although encouraging, the Lancioni et al.’s (2015) data can only be viewed as preliminary. Additional participants should be exposed to the program to determine its overall robustness and dependability in supporting exercise and positive personal involvement (Kazdin, 2011). Moreover, the program might be focused on supporting a leg-raising response (a most basic/direct exercise movement for persons who are sedentary and do not use their legs; M. Brown et al., 2000; Eggermont et al., 2010; Harvey, Smith, & Jones, 1999) rather than the arm-raising and leg-pedaling movements used earlier. In line with the aforementioned points, the present study extended the early assessment of the program by (a) involving nine new participants with advanced Alzheimer’s disease and (b) requiring the participants to exercise a leg-raising response such as the one referred to above (Kazdin, 2011). The study also checked whether the program had an impact on participants’ signs of positive personal involvement (Dillon & Carr, 2007; Lancioni, Singh, O’Reilly, Green, et al., 2013).
Method
Participants
The participants included eight women (Participants 1-8) and one man (Participant 9), who were between 75 and 91 (M = 84) years of age and represented a convenience sample (Pedhazur & Schmelkin, 1991). They lived in medical care centers for people with Alzheimer’s disease and other dementias. Medical and psychological records indicated that they were in the low moderate or severe level of Alzheimer’s disease, with scores of 7 to 13 (M = 9.7) on the Mini-Mental State Examination (Folstein, Folstein, & McHugh, 1975). Participants 5, 6, and 8 received pharmacological treatment for the Alzheimer’s condition in the form of memantine. Participants 1, 2, 4, and 7 received antipsychotic or antidepressant medication. Their selection for the study was based on the criteria used by Lancioni et al. (2015). That is, they were sedentary and generally passive (i.e., did not engage in social interactions or functional activities/movements). Yet, all of them could perform the leg-raising response, which was targeted in this study (and deemed useful for their condition; see Eggermont et al., 2010; Hoffmann et al., 2013; Pareja-Galeano, Garatachea, & Lucia, 2015; Rolland et al., 2008). They also showed clear interest in a number of environmental stimulation events (e.g., music and videos) and the ability to respond to verbal reminders/prompts directed at helping them perform small arm or leg movements. Finally, staff and families (a) had expressed their support for a technology-aided program that would promote a leg response through the delivery of preferred stimulation contingent on that response and prompts in case of passivity, and (b) believed that such a program could be beneficial and enjoyable (given the stimulation opportunities available) for the participants (Brett, Traynor, & Stapley, 2016; Ferrero-Arias et al., 2011; Halpern, 2012; Letts et al., 2011). The participants’ families had provided informed consent for this study, which had been approved by an Italian ethics committee.
Sessions, Leg Response, Technology, Stimulation, and Data Recording
Sessions lasted 5 min and were carried out two to six times a day, on an individual basis, in quiet areas of the centers in which the participants lived. The leg-raising response consisted of lifting the left or right leg (i.e., always the same one or interchangeably). The technology included a microswitch, a computer with sound amplifier, and basic software. The microswitch consisted of a tilt device or a combination of two such devices fixed onto the participant’s leg(s) and served to monitor the leg response (Lancioni, Sigafoos, O’Reilly, & Singh, 2013). The computer’s single function during the baseline was to record the leg responses. The computer’s functions during the intervention included (a) delivering a 10-s stimulation period after each leg response, (b) presenting a verbal prompt (reminding/encouraging the participants to respond) after periods of about 15 s of no responding, and (c) recording both the leg responses and the prompts.
The stimulation consisted of 10 s of old songs, prayers, and religious hymns or videos of specific activities or events considered preferred for the participants based on staff reports and direct screening. Screening involved 10 or more nonconsecutive presentations of brief segments of several stimuli. A stimulus was selected for use during the study only if the two research assistants conducting the screening agreed that it produced positive reactions (e.g., orienting, smiling or positive verbalizations) in at least 60% of the presentations (Lancioni, Singh, O’Reilly, Sigafoos, et al., 2013; Staal, Pinkney, & Roane, 2003).
Data recording also involved participants’ signs of positive personal involvement (i.e., singing, music-related body movements, positive verbalizations, and smiles; Lancioni, Singh, O’Reilly, Green, et al., 2013). This recording was carried out during the baseline sessions and 18 to 20 intervention sessions spread over the intervention period for all participants except Participant 7 who left the program prematurely (see below). For such recording, the research assistants relied on videos of the sessions. Recording was carried out through a partial interval system, in which 10-s observation intervals were followed by 5-s scoring periods (Kazdin, 2001). Interrater agreement (computed over 10 sessions per participant by dividing intervals with the same scoring by the total number of intervals observed and multiplying by 100) ranged between 75% and 100%, with individual means exceeding 90%.
Experimental Conditions and Data Analysis
The study was carried out according to a non-concurrent multiple baseline design across participants (Barlow, Nock, & Hersen, 2009). The baseline phase included two or three sessions for Participants 1, 2, 6, and 7, and four or five sessions for the other participants. The number of sessions was decided prior to the start of the study with the stipulation that a participant would be provided with extra sessions if his or her response frequency was above three and presented an increasing trend. (This condition never occurred.) The intervention phase included 43 sessions for Participant 7 and between 92 and 149 (M = 120) sessions for the other participants. The low number of sessions for Participant 7 was due to her being transferred out of the medical care center in which she lived. During the sessions, the participants sat in their wheelchairs or regular chair and had the technology, which was set up in advance by the research assistants in charge of the study. Prior to the start of each baseline session and each of the initial 25 to 40 intervention sessions, the research assistants also guided the participants to perform the target leg response. Changes in leg responses and signs of positive personal involvement (see above) from the baseline to the intervention phase were analyzed via the “Nonoverlap of All Pairs” (NAP) method (Parker & Vannest, 2009; Parker, Vannest, & Davis, 2011).
Baseline
During the baseline sessions, the technology recorded the participants’ leg responses but did not provide stimulation or prompts.
Intervention
During the intervention sessions, the technology provided stimulation and prompts while recording both responses and prompts (see above). The intervention sessions were preceded by three or four practice sessions, in which the research assistants used the necessary guidance to ensure participants’ responding to the computer-delivered verbal prompts and experience of the stimulation events available for the leg response.
Results
Figures 1 and 2 summarize the participants’ baseline and intervention data for leg responses and prompts (i.e., by plotting those data over bocks of sessions). The five panels of Figure 1 summarize the data for Participants 1 to 5, respectively. The four panels of Figure 2 summarize the data for Participants 6 to 9, respectively. Each bar and white circle combination represents the mean frequencies of leg responses performed and of prompts received per session, respectively, over a block of sessions. The number of sessions included in each block is indicated by the numeral above it.

The five panels summarize the data for Participants 1 to 5, respectively. Each bar and white circle combination represents the mean frequencies of leg responses performed and of prompts received per session, respectively, over a block of sessions. The number of sessions included in each block is indicated by the numeral above it. The ordinate values vary across participants.

The four panels summarize the data for Participants 6 to 9, respectively. Data are plotted as in Figure 1.
During baseline, the mean frequencies of leg responses ranged from zero to slightly above two per session. Prompts were not available; thus, their frequencies were reported as zero in the figures. During the intervention (carried out over periods of nearly 2 to more than 3 months except in the case of Participant 7 who attended for less than 1 month), leg responses increased. Their mean frequencies ranged from nearly 10 (Participant 1) to nearly 17 (Participant 2) per session, with an overall mean of about 13 per session. Pairwise comparisons across the data points of the baseline and intervention phases (according to the NAP method) yielded NAP and Taunovlap indices of 1.0 for all participants (see Parker et al., 2011). Indeed, all intervention data points were higher (showed improvement) compared with the baseline data points, suggesting that the intervention (preceded by the three or four practice/introductory sessions) was highly effective in enhancing performance (Byiers, Reichle, & Symons, 2012; Parker et al., 2011). The mean frequencies of verbal prompts provided by the computer varied from about two (Participants 2) to more than seven (Participant 1) per session. The mean percentages of intervals with signs of positive personal involvement were between zero and 10 during the baseline sessions and increased to between about 35 (Participant 8) and 55 (Participant 6), with an overall mean of about 45, during the 18 to 20 intervention sessions in which such measure was recorded. This measure was not available for Participant 7 (see above). Pairwise comparisons across the data points of the baseline and intervention phases (according to the NAP method) yielded NAP indices of .98 for Participants 1 and 8 and of 1.0 for the other six participants for whom the measure was recorded. The Taunovlap indices were .96 and .97 for Participants 1 and 8, respectively, and 1.0 for the other six participants (Parker et al., 2011).
Discussion
The results indicate that the participants had a clear improvement in both leg responses and positive personal involvement during the intervention phase of the study. In practice, a simple technology-aided program was effective in helping participants in the moderate to severe level of Alzheimer’s disease to engage in mild physical exercise, that is, to independently perform a leg-raising response recommended for individuals who are sedentary and do not use their legs (M. Brown et al., 2000; Eggermont et al., 2010; Harvey et al., 1999; Pareja-Galeano et al., 2015; Wang & Spillane, 2009). The participants also seemed to enjoy their exercise sessions, showing extensive signs of positive personal involvement during those sessions (Dillon & Carr, 2007; Godwin, Mills, Anderson, & Kunik, 2013; Lancioni, Singh, O’Reilly, Sigafoos, et al., 2013). These results, which are consistent with pilot findings in this area (Lancioni et al., 2015), may lead to a number of considerations.
First, in light of the data, one might suggest that a new practical exercise option is now available for persons with advanced Alzheimer’s disease who are sedentary and largely passive throughout the day. Indeed, the technology-aided program assessed in this study might represent a fairly convenient approach for those persons and support their successful exercise engagement. Such an engagement (a) is considered beneficial/desirable for their physical condition and their social image (Eggermont et al., 2010; Hernández et al., 2015; Perales, Cosco, Stephan, Haro, & Brayne, 2013; Williams & Tappen, 2008; Wood, Womack, & Hooper, 2009; Zhu et al., 2015) and (b) appears obviously relevant for their emotional well-being and involvement (Brett et al., 2016; Friedman, Wamsley, Liebel, Saad, & Eggert, 2009; Lancioni, Singh, O’Reilly, Sigafoos, et al., 2009; Lancioni, Singh, O’Reilly, Zonno, et al., 2009). It is also noteworthy that such an engagement might be achieved with a relatively small cost in terms of staff’s time and commitment, and thus might be viewed as practically affordable (Lancioni et al., 2015; Staedtler & Nunez, 2015).
Second, the use of stimulation contingent on the leg response during the intervention phase of the study may have been essential to motivate the exercise engagement of all participants (Catania, 2012; Kazdin, 2001). The role of the prompts was seemingly different across participants. In fact, some of the participants produced relatively high response frequencies with reasonably low levels of prompting (i.e., Participants 2, 3, and 6). Some of these participants might have remained fairly busy even without systematic prompting (Kazdin, 2001; Pierce & Cheney, 2008). Other participants, on the contrary, appeared to clearly depend on the prompts to maintain a relatively consistent response level (i.e., Participants 1, 4, and 8). For the latter participants, the stimulation, albeit ostensibly pleasing, seemed to be insufficient to guarantee performance continuity. Whether the medication used by seven of the participants had any effects on their overall level of prompt reliance cannot be determined on the basis of the data available. However, it seems quite clear that medication was not responsible for the change observed from the baseline to the intervention. In fact, medication was present before the study and throughout its implementation phases.
Third, the participants’ increased signs of positive involvement during the intervention sessions might be taken as an indirect indication of (a) the acceptability of the response exercise required during the sessions and (b) the apparent enjoyableness of the procedural conditions available (i.e., of the stimulation delivered contingent on the responses). Providing the participants with an acceptable and enjoyable engagement situation might be instrumental for (a) improving their status and quality of life (Brett et al., 2016; R. I. Brown, Schalock, & Brown, 2009; Letts et al., 2011; Pilotto et al., 2011; Sunderland, Catalano, & Kendall, 2009) and (b) increasing their motivation to participate in the sessions and remain active (Noguchi, Kawano, & Yamanaka, 2013; Pierce & Cheney, 2008). Obviously, performance maintenance would depend on (a) the ability of the care context to use a variety of stimuli so as to avoid a decline in stimulation impact and (b) the stability of the participant’s situation or adjustments/facilitations of the responses required of them (Pierce & Cheney, 2008; Soto et al., 2012; Wilson et al., 2012). With regard to this point, it might be noted that the participants of this study maintained stable responding, without any adjustment needs, throughout the intervention period.
Fourth, the technology used in the study is quite simple. In fact, it includes a basic microswitch to monitor the participants’ target response and a computer to deliver contingent stimulation and prompts. The same simple technology, with possible changes or adjustments in terms of microswitch, might be employed for a variety of participants and for a number of arm and leg responses, such as arms sideward stretching and leg-pedaling movements performed from a sitting position (Lancioni et al., 2015). The simplicity, practicality, and alleged reliability of the technology, together with its rather limited cost (estimated to be around or below US$1000), could be important elements in determining its general affordability as well as its definite adoption and continued use in daily contexts dealing with persons with advanced Alzheimer’s disease (Baxter, Enderby, Evans, & Judge, 2012; Dahlin & Rydén, 2011; Fried-Oken, Beukelman, & Hux, 2012; Lenker, Harris, Taugher, & Smith, 2013; Scherer, Craddock, & Mackeogh, 2011).
In conclusion, the results of this study are quite encouraging, but caution is required in making wide-ranging statements given the early stages of this type of research with participants with Alzheimer’s disease and the relatively small number of participants so far exposed to the program. New research should extend the assessment of the program with additional participants to (a) determine the robustness of the present procedural conditions and the reliability of the technology solutions adopted, (b) gain estimates of duration/maintenance of the performance and mood improvements, and, eventually, (c) remedy obvious limitations of the present study. New research would also need to target additional responses (e.g., leg or arm sideward movements) to identify the effectiveness of the procedural conditions and technology solutions available in supporting those responses. A final research point could concern the acquisition of staff and families’ suggestions for the improvement of extant program components (Barlow et al., 2009; Borg, Larsson, & Östergren, 2011; Callahan, Henson, & Cowan, 2008; De Joode, van Boxtel, Verhey, & van Heugten, 2012; De Joode, van Heugten, Verhey, & van Boxtel, 2013; Melnyk, 2012).
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
