**Acknowledgements**

During each effort, the muscle activity of TA, GAS, RF and BF were collected at 2000Hz with a time constant of 0.03sec equal to 5.3Hz high-pass filter. After the collection, the data was filtered with low and high pass filter of 500Hz and 10Hz, respectively. Then, the data was fullwave rectified and low-pass filtered with moving average at 5 Hz to obtain the linear envelope. Following this, the data was normalized from heel contact (0%) to next heel contact (100%), and expressed as %MVC. The comparison was then made between each water-walking speed

A normalized time course pattern and results of the cross correlation analysis are shown in Figure 22. The cross correlation coefficients were moderate in the RF and BF (r = 0.53 - 0.70), and were high in the TA and GAS (r = 0.83 - 0.90). This showed the muscle activity pattern of the water-walking were similar to that of the land-walking even in the slow and fast speed. Moreover, as seen in the figure 22, the muscle activation was higher during the water-walking than land-walking during most part of one full stride for all muscles except the GAS. This suggests the water-walking would be able to simulate the land-walking very closely regardless of the speed of the water-walking, and stimulate the thigh muscles and TA sufficiently even in the slow speed. The authors also suggest that even during slow speed, water-waking is an effective exercise modality for muscle training in a similar way to normal walking on land.

A possible limitation can be seen when applying normalization method to time course pattern in that the normalization process resulted in apparent cancelations in time length changes which may alter the sequencing and timing of events especially in exercise in water due to the buoyancy and water resistance affect on the movement duration. In addition, changing the time length may also distort the time course pattern, where the exact timing may not be comparable between the same timing of two wave forms (i.e. does the 50% of land-walking truly match to the 50% of water-walking in normalized data?). This should be considered in

This chapter focused on muscle activity during exercises in water especially variations of gait (walking and running) and other activities of daily living. Descriptions on how to waterproof electrodes, placing the electrodes on the muscle of interest, and the verification of EMG collection in the water environment was then discussed. The authors also suggest that when the electrodes are waterproofed appropriately, EMG recording is not affected by water immersion. Namely, we can measure muscle activity correctly even in underwater condition

In summary of the reported characteristics of muscles activity during walking in water, TA and thigh muscles and ES tended to show higher activity than land-walking with self-selected pace. There was a lower activity than land-walking with Triceps Surae muscle, VL and abdominal muscle. Further, most muscles tended to decrease their activity when walking on treadmill apparatus, compared to land treadmill walking. During DWR the characteristics of muscle activity for the thigh, hip and trunk muscles show higher activity than walking in water and/or on land, whilst the muscle activity of Triceps Surae muscle decrease dramatically. From these

and confortable speed of the land-walking.

232 Electrodiagnosis in New Frontiers of Clinical Research

studying outputs from EMG modality of water-walking.

**7. Conclusion**

without any artifact.

The authors thank PhD Hitoshi Wakabayashi in Chiba Institute of Technology, PhD Daisuke Sato in Niigata University of Health and Welfare, and Professor Takeo Nomura in NPO Tsukuba Aqua Life Research Institute for your contributing to the data collection related in this chapter.
