This study aims to examine the effect of external, internal and no attention focus walking training during gait rehabilitation on real-time conscious motor processing (reinvestment), balance, walking ability and fear of falling by older adults at risk of falling in Hong Kong.

One-hundred and eight older adults will be recruited from elderly community centers in Hong Kong. Participants will be randomly assigned into 3 groups (i.e., No Attention Focus Walking Group (NAFWG; active control group, n=36), an External Attention Focus Walking Group (EAFWG, n = 36) or an Internal Attention Focus Walking Group (IAFWG, n = 36)). Participants in different groups will have training sessions (about 45 minutes each) three times per week for 4 weeks in a group of 6 participants. A total of 12 sessions will be completed by each participant.

All training sessions will be conducted by experienced registered physiotherapists in Hong Kong and a research assistant with experience in exercise training for older adults. In each training session, all groups will have warm-up (5 minutes), balance training (5 minutes), body transport training (5 minutes), body transport with hand manipulation training (5 minutes), walking training with various levels of difficulties in a 40-meter walkway with different instructions in different walking groups (20 minutes) and cool down (5 minutes). For the walking training (20 minutes), all participants will be invited to conduct walking training on a walking field with an area of 25 meter square and a total walking distance of about 40 meters for each walking trial from cone 1 to 9. Two screens that connected with a laptop computer will be positioned 1 meter beside the walking field. Both screens will be projected different digits from 0 to 9 randomly in the speed of 2 seconds per digit. Participants in the NAFWG, EAFWG, and IAFWG will receive different instructions during walking training.

Each participant will complete assessment sessions (total 3 assessment sessions) before training at baseline (T0), just after completion of all training sessions (T1) and 6 months after completion of all training sessions (T2). In the baseline assessment (T0), a structural questionnaire will be used to ask for demographics, medical history, detailed history of fall incident, social history and social economic status of all participants. A battery of assessments will be conducted to assess physical and cognitive abilities of the participants in all assessment sessions (T0, T1, & T2). Walking ability will be assessed by the 10 meters comfortable and fast walking speed (Bohannon, 1997). Functional balance and gait assessment will be done by the Tinetti Balance Assessment Tool (Tinetti, 1986), the Berg Balance Scale (BBS) (Berg et al., 1989) and the Timed 'Up & Go' Tests (TU&G) (Podsiadlo & Richardson, 1991). Cognitive function will be evaluated by the Chinese version Mini-Mental State Examination (MMSE-C) (Folstein et al., 1975; Chiu et al., 1994). The Chinese version of the Fall Efficacy Scale International (FES-I (Ch)) (Kwan, Tsang, Close & Lord, 2013) will be completed to assess the fear of falling. The Chinese version Movement Specific Reinvestment Scale (MSRS-C) (Masters et al., 2005; Wong et al., 2015a; Wong et al., 2015b) will be administered to examine the conscious motor processing propensity (i.e., movement specific reinvestment). The alpha2 EEG coherence between T3 (verbal-analytical region of the brain) and the Fz (motor planning region of the brain) (i.e., T3-Fz EEG coherence) of all participants when walking at the 6-meter level-ground walkway (three walking trials) will be determined to identify the real-time conscious motor processing propensity (Zhu et al., 2011; Ellmers et al., 2016; Chu & Wong, 2019). All participants will be equipped with EEG electrodes before the start of the three walking trials. EEG activity will be received using a wireless EEG device (Brainquiry PET 4.0, Brainquiry, The Netherlands) and will be recorded using the real-time biophysical data acquisition software (BioExplorer 1.5, CyberEvolution, US). Previous research has demonstrated that alpha2 (10-12Hz) T3-Fz EEG coherence is sensitive at detecting within-subject changes in real-time conscious motor processing propensity during a postural sway task (Ellmers et al., 2016). T4-Fz EEG coherence will be utilized to identify whether the changes in the alpha2 T3-Fz EEG coherence will be due to global activation of the brain. The EEG electrodes are non-invasive and will not be used in any diagnostic purpose.

All participants will be asked to record their number of falls prospectively at the time between T1 (completion of all training sessions) and T2 (6 months after completion of all training sessions) using a structural calendar. The number of falls within the 6-month follow-up period will then be collected.

Inclusion Criteria:

• Age 65 or above;
• No history of cerebral vascular disease, Parkinson's disease or other neurological deficit;
• Chinese version of the Mini-Mental State Examination (MMSE-C) total score of equal or more than 24;
• Able to walk independently indoor for at least 40 meters continuously with comfortable pace;
• Older adults with moderate to high risk of falling, as indicated by the score of less than 24 out of 28 in the Tinetti Balance Assessment Tool.

Exclusion Criteria:

• Any potential participant who cannot meet the inclusion criteria.