IMPORTANCE OF EXPERIMENT DESIGN
Every scientific research study starts with a question and ends with a possible solution. In experimental research, especially in brain research, the question may be as general as “what brain regions are associated with the state of depression and stress?” or “what is the role of alpha band in stress and anxiety?” The experimental question pushes the researchers to derive a research hypothesis, which is a description about how a given manipulation can change certain measurements. The research hypothesis may be general or specific, like a research question. An example of research hypothesis in EEG research is the statement, “EEG alpha activity at frontal cortex will be reduced in depression patients.” In a hypothesis statement, the researchers can make particular claims, such as, “EEG frontal alpha activation in right frontal region will be high compared to left frontal lobe.” The scientific characteristic of a hypothesis is that it should be tested experimentally. Highly specific hypotheses can be easily falsified and are more informative.
To test a hypothesis, a researcher or experimenter designs an experiment. In experiments, first manipulate some aspect of the research problem and then measure the outcome of that manipulation; e.g., Galileo’s experiment of the effects of mass on gravity. He speculated that an object’s acceleration due to gravity is independent of its mass. He tested this hypothesis by dropping two balls of different masses from a certain height and observed that they fell at an equal rate. In this procedure, first he manipulated the mass of the objects being dropped and then measured the time required for the balls to fall from a given height. The manner in which a researcher organizes the manipulations and measurements of an experiment is the experimental design. The required skills for experimental design may differ from one discipline to another; however, in this chapter the focus is on EEG experiment design.
In general, every researcher wants to investigate their research question about the world and design an experiment to find a meaningful solution to this research question in the most efficient way. Thus, the experiment needs to be designed well in each and every aspect.
In EEG research, significant resources are required in terms of equipment cost, time, and human resources (including experimenters, subjects, and research assistants). The EEG setup procedure normally takes 20-30 minutes; data collection sessions vary widely depending on the number of trials or conditions. Typical EEG experiments have up to 20 participants per condition or group of interest.12 The number of participants varies based on the size of the effect to be tested and the number of trials to be collected per participants. In addition, the raw EEG requires human efforts to clean the unwanted artifacts present in the signal. Therefore, an inadequate EEG experiment design will create great trouble to the researchers, because it will either fail to answer the defined hypothesis or provide results that are hard to interpret to make a conclusion. Thus, all the investments in the form of time and money would be wasted. The researchers are required to make a good plan and consider each and every step of the experiment from participant recruitment to interpretation of the final results, including all the possible risks, restrictions, confounding variables, and resources. The characteristics of well-designed experiments are as follows13-15:
- • as simple as possible to operate for the experimenter and easy to reproduce for later researchers
- • tests a specific hypothesis and provides fair estimates of the factor effects and associated risks
- • minimum cost of running experiment and enables experimenter to detect significant differences
- • includes planning for data analysis and results interpretation
- • allows to make conclusions that have wide validity
These characteristics are very general; each researcher is required to look the associated different parameters with respect to the research problem and the available resources in order to make a well-designed EEG experiment. For particular guidelines for subject preparation in EEG and ERP experiments with human participants, see the work of Light and colleagues.12