Modern medicine would be unthinkable without the possibility of administering pharmaceuticals and other evidence-based interventions. The development of these interventions requires scientific research, ultimately with human subjects. This venture raises ethical, legal and human rights issues, which are addressed in numerous national and international declarations and regulations. In these documents, special attention is usually directed towards research involving vulnerable groups, such as children, pregnant women, unemployed persons, refugees, patients with psychiatric disorders, dementia or mental retardation, and those who are dying. In relation to patients with psychiatric disorders, two important and mutually connected ethical questions can be posed. Firstly, is research with persons who have severe psychiatric illnesses permissible? And, secondly, how can the mental capacity of prospective research subjects be assessed? We investigate these questions using the Dutch legal system as an example. Regarding the first question, the Dutch Medical-Scientific Research on Human Subjects Act (1998) presents a detailed regulation that is in line with relevant international documents, such as the Convention on Human Rights and Biomedicine (1997). In the Dutch statute, the possibilities for research involving subjects who lack mental capacity are limited, but not completely excluded. Under certain conditions, two types of research are exempted from the general prohibition of research with such subjects that is included in article 4 of the statute. These two types are (i) therapeutic research and (ii) non-therapeutic research that could not take place without the participation of subjects from the category to which the mentally incapacitated person belongs. The conditions pertain to ethical and scientific review, insurance, written proxy consent and respect for resistance by the subject. An extra condition for the permissibility of non-therapeutic research is that the risks for the prospective subject are negligible and the burdens minimal. Although the Dutch regulation obviously does not solve all problems, it is relatively clear when compared with the situation in other European countries, such as Belgium, France, Germany, and England and Wales. Regarding the second question, two basic factors need to be considered when defining 'mental capacity'. These relate to the assumption of competence and to the task-specificity of capacity. The crucial issue in assessing mental capacity is not whether a psychiatric diagnosis is present, but whether the patient has the mental abilities required to make the decision at hand in a meaningful way. In establishing an appropriate standard for capacity assessment, several interests have to be weighed. The ethical demands of protection of subjects and stimulation of scientific research may be balanced by attempting to enhance patients' mental capacity. The procedure of 'experienced consent' seems promising in this regard, although this approach entails its own ethical problems.