1

1 Nature of the interface
Biosensor, in essence, is a device that detects and measures electric impulses in biological elements such as skin, muscle or brains. Biosensors can also be used to detect certain biological analytes or in other words, components, from chemical substances. This could be the concentration of a distinct conductive element in the blood flow of a person for example or a toxin in a food product. The use of biosensors provides a multitude of different inputs for a user interface. Biosensors themselves are rather inadequate in order to be used as the sole way to control a user interface. Instead they can be used to enhance the user experience and widen the range of input methods used in an interface. For example the use of biosensors enables a system to respond to the user’s heartbeat rate and muscle tension or to identify the position and movement of body parts such as fingers. Currently, existing biosensors measure physiological activity, muscle electrical activity, brain electrical activity, and eye movement among other things. Extracting accurate physiological data from biosensors is often a complex task. In particular, extracting data from different typologies of biosensors will require architecture of great flexibility and the possibility to connect them to different external monitoring devices. Biosensors are key components in both physiological and psychophysiological computing. Psychophysiological computing can be considered to be the same as affective computing, where human-computer interactions are based on the users affects. The field of affective computing involves computer science, psychology and cognitive science. Physiological computing can be seen as more straightforward version of affective computing. Instead of human-computer interactions, one could refer to muscle-computer interactions when describing physiological computing. In this field, biosensors have a much more direct effect on the inputs to the interface than in affective computing, where a multitude of direct inputs are interpreted as a whole to determine the users emotions for example. This paper will concentrate on physiological computing as this field is more directly related to biosensors in user interfaces.

2 Technology
Biosensors can be roughly divided into two categories: sensors for biochemical analyte detection and sensors for detecting electrical activity in biological substances. Strictly speaking, only the aforementioned category fills the textbook definition of a biosensor, but sensors belonging to the second category, such as heart-rate, brain activity and muscle activity sensors, which measure electrical activity, are generally recognized as biosensors [1]. The technology behind both types of sensors will be inspected next.

2

2.1 Analyte detection
Biosensor technology couples knowledge of biology with advances in microelectronics. A biosensor is an analytical device for the detection of a specific analyte that combines a sensitive biological component, such as a cell or an antibody, with a physicochemical detector component. [1]. A biosensor for analyte detection, illustrated in figure 1, consists of three components [2]:  Sensitive biological element  Detector  Transducer

Figure 1. The structure of a biosensor

Biosensors are detecting devices that rely on the specificity of cells and molecules to identify and measure substances at extremely low concentrations. When the substance of interest collides with the biological component, the transducer produces a digital electronic signal proportional to the concentration of the substance. [3] This kind of biosensors can measure the nutritional value, freshness and safety of food, provide emergency room physicians with bedside measures of vital blood components, and locate and measure environmental pollutants. [3]

2.2 Electrical activity detection
2.2.1 Electromyography Biosensors used in electromyography are a neural interface technology that detect nerve and muscle activity. Among the detectable and measurable activities are muscle electrical activity, brain electrical activity, and eye movement. Biosensors are electrodes that sit on the skin over the 3

muscle or nerve being sampled. Eye movement, for example, is determined from biosensors placed strategically on the forehead and under the eyes. [4] Electrical signals have many measurable qualities, including intensity and spectral characteristics. Energy is also measurable from a multitude of motor units. Just as the brain uses these signals to control functions of the human body, these signals can be detected by biosensors and then interpreted by software to control electronic devices external to the human body. [4] 2.2.2 Electrodermal acitivity Electrodermal activity is electrical activity of the sweat glands in the skin. Tests of electrodermal activity have indicated correlation between skin conductivity and tendency to delinquency, aggressiveness and recidivism. Lie detectors or polygraph tests use electrical skin conductivity and heat rate to measure responses to questions. [5]

3 Current implementations
3.1 Analyte detection
Traditional biosensors, which detect analytes aren’t commonly used in user interfaces, as their main purpose is to detect and measure the concentrations of chemical substances. Commercially successful implementations of these biosensors are generally from the field of health and medicine, where the historically most important application has been blood glucose level monitoring for diabetes patients. Other uses for such biosensors have been found in the food industry among other industries where the consistency of biochemical substances is of importance. [6] Other, more human related implementations of traditional biosensors are Body Sensor Networks, or BSN’s. Such networks incorporate a multitude of biosensors attached to a person in order to monitor the health and wellbeing of the user for example. BSN’s have been implemented in the monitoring of the vital signs of chronically ill persons. An example would be an automated insulin injector for a diabetes patient. Other applications can be found from the field of sports and military, where biosensor data is monitored to determine the performance level of a person. [7]

3.2 Muscle-computer interaction
An extremely interesting field of biosensor study is the field of muscle-computer interaction. By measuring the muscle activity of the user’s arms, intricate hand gestures and even the amount of pressure applied can be used as an input without the use of any tactile input devices. These features can be used to enhance the usability of touch screen interfaces by identifying which 4

finger is actually touching the screen and the pressure applied to the screen. Hand gestures such as pinching can be used to enable completely non-touch interactions. A device that can be used for such muscle-computer interactions developed by Saponas et al. is illustrated in figure 2. The latest version of device is actually just an armband that is placed over the forearm of the user. The data measured by the sensor system is transmitted wirelessly. [8]

Figure 2. Muscle-computer interface [8]

Another interesting application is the mouth-computer interface illustrated in figure 3. While the actual inputs are received from a group of optical sensors manipulated by the users tongue, the device could be enhanced with analyte detective biosensors in order to receive real time analysis of the food that the user consumes. Users jaw tension and movement could be measured with biosensors that detect electrical activity.

Figure 3. Mouth-Computer Interfaces

Biosensors are also used in a bionic contact lens development project, where the goal is to come up with a contact lens that can be used as a head-up display, or a HUD, and an input device which provides many possibilities for the implementation of augmented realities. Integrated biosensors 5

can be also used to measure some of the user’s biological properties, such as blood glucose level straight from the user’s eye. A schematic of such a contact lens is illustrated in figure 4.

Figure 4. Bionic contact lens

3.3 Gaming
Some implementations can be found from the field of computer gaming. A biathlon game involving cross country skiing and target shooting developed by Nenonen et al. in Helsinki University of Technology applied the user’s heart rate as the main input to the skiing part of the game. The skiing speed of the avatar in the game was directly related to the user’s heart rate. During the shooting phase of the game, the user’s heart rate affected the stability of the aiming. During the skiing part of the game, the user had to either paddle an exercise bicycle or a mini stepper to control their heart rate. They were also equipped with two wands with a button that emulated the ski poles. Other inputs that can be used in the gaming environment are facial expressions and skin conductivity. These two inputs alongside heart rate reflect the stress and excitement levels of a gamer [9][10]

4 Human factors and usability issues
An application equipped with biosensors requires the user to wear some kind of a device that contains the necessary sensors. Such a device could for example be a glove, a suit or a helmet. This has direct implications on the usability of such interfaces. In order to measure health parameters, biosensors must be in close contact with the skin, and sometimes, even inside the human body. Several challenges are related to this kind of measuring are identified, among them are mentioned limited sensor power, limited computational power, material constraints, continuous operation, robustness and fault tolerance, scalability, security and interference, and regulatory requirements. Moreover, the choice of hardware must be careful, namely, in terms of size for non- intrusive wear, produced heat, and battery life. [7]

6

The biosensors cannot accurately interpret muscle activity straight away. Software must be trained to associate the electrical signals with different gestures. This is an obvious usability issue as any user interface controlled by biosensor input can’t be designed as a walk-up-and-use system. To ensure correct action to input relationships, the training of the software has to be quite extensive and error free, as the user is unable to affect the inputs to the system after the controls have been mapped to user actions. [11]

5 Testing techniques
As was earlier stated in this paper, extracting accurate physiological data from biosensors is often a complex task. When the data that is received from the biosensors is used to create direct interactions between the source of the data, a muscle for example, and a computer system, the accuracy of the mapping of the source data to correct actions becomes the most critical factor for success. In such systems tests that provide quantitative data are preferred, such as the ratio of successful actions to correct activity or the accuracy of the input to the system [8]. In more subjective solutions where biosensors are used to enhance the user experience, qualitative measures are preferred. Such a testing measure could be user satisfaction, which is an appropriate measure for gaming, where user interactions are mainly driven by the desire to be entertained [12].

6 Application case study
Heureka, The Finnish Science Centre, developed a game that utilizes biosensor technology for their “Move and Play!” exhibition. The concept of the game is to use heart rate as the sole input to a simple game, where the user has to guide a red blood cell through tunnel-like levels that depict the cardiovascular system of a human. A screenshot of the game is illustrated in figure 5. The input to the game can be received from any generic heart rate monitoring device. Initial tests were done using a Suunto transmitter belt and a PC Pod, which receives the heart rate data from the belt and relays it forward to the PC. The goal of the game is to move the red blood cell through the blood vein within the given timeframe. The cell is controlled by the user’s heart rate in such way, that an increase in the heart rate moves the cell upwards and forwards and a decrease in the heart rate moves the cell downwards and forwards. Keeping the heart rate level stops the cell from moving anywhere. The characteristics of the environment, as in the blood vein can be adjusted to create a different gaming experience. Adjustable parameters include the friction of the walls of the blood vein and the elasticity of the red blood cell among with factors that affect the rate of the games response to the changes in the heart rate. 7

Figure 5. Screenshot of the heart rate game

The purpose of the game was to introduce users to a new way of interactive controlling of games and to encourage gamers to incorporate exercise to their gaming routines. Similar goals have been set for the Nintendo Wii’s Wii Fit game and it has been met with extensive interest from the public.

7 Future trends
The ongoing development and hype around virtual reality and immersive environments suggest that those fields will continue to develop and gain ground as a new way to interact with devices. Control devices based on biosensors, such as the muscle-computer interaction armband presented in this paper can offer a multitude of intuitive and natural ways for user’s to interact with these new environments without the use of intrusive and unintuitive tangible controllers. The area of telemedicine and automated healthcare is another field where the use of biosensors in BSN’s is highly sought after and still in a very early phase of development. Visions of the future of this field are quite plausible and the healthcare industry is remarkably large, which means that the development in this field will be rapid. Gaming industry is experiencing a paradigm shift from traditional controlling devices to new more interactive ways to control games, with Nintento Wii in the forefront of this new wave. Motion detection with accelerometers and cameras is already a commercialized way to control games and the industry is looking for new ways to further enhance the gaming experience. The musclecomputer interface developed by Saponas et al. has already been used as a controller for the popular Guitar Hero video game, where gamers use a guitar shaped controller to “play”. The interface allows the gamer to play the game without the tangible controller by just using their fingers for pinching and strumming. [8] 8

8 References
[1] Kress-Rogers, E. Handbook of Biosensors and Electronic Noses: Medicine, Food, and the Environment. 1st. CRC Press, Inc. (1996) [2] Knotts, T. Biosensor Basics. URL:http://www.sas.aamu.edu/physics/events/workshop/2007/workshopfiles/RISEJul07Basics.pdf. Date referred 3.11.2010. [3] Scienceniche.com. Biosensor technology. URL:http://scienceniche.com/applied_science-_technologies/technology/biosensortechnology.html. Date referred 3.11.2010. [4] Rajmohan, J. Biosensors. Isha Books, India (2006) [5] Malmivuo, J & Plonsey, R. Bioelectromagnetism. Oxford University Press (1995) [6] Newman, J & Turner, A. Home blood glucose biosensors: a commercial perspective. Biosensors and Bioelectronics. Volume 20, Issue 12, 15 June 2005 [7] Rodrigues JJPC, et al. Biofeedback data visualization for body sensor networks. J Network Comput Appl (2010) [8] Saponas, S, et al. Enabling Always-Available Input with Muscle-Computer Interfaces. Proceedings of ACM UIST 2009 (2009) [9] Nenonen, V et al. Using Heart Rate to Control an Interactive Game. CHI 2007 Proceedings (2007) [11] Technologyreview.com. Muscle-Bound Computer Interface - Technology Review. URL:http://www.technologyreview.com/computing/23813/?a=f. Date referred 3.11.2010 [12] Nacke, L et al. More than a feeling: Measurement of sonic user experience and psychophysiology in a first-person shooter game. Interacting with Computers Volume 22, Issue 5 (2010)

9