N.P. Spanos et al begin their article with a neat encapsulation of the status of psychological research into glossolalia:
"Glossolalia (i.e., speaking in tongues) is vocalization that sounds languagelike but is devoid of semantic meaning or syntax. In the Christian tradition this vocalization pattern is associated with the ideas of possession by the Holy Spirit and communication with God through prayer or prophecy. Some scientific investigators conceptualize glossolalia as the product of an altered or dissociated state of consciousness, whereas others view it as symptomatic of psychopathology.
"The available empirical data fail to support either of these hypotheses. For example, both ethnographic observations and experimental findings indicate that glossolalia can occur in the absence of kinetic activity, disorientation, and other purported indexes of trance, and that experienced glossolalics do not differ from nonglossolalic controls on measures of absorption in subjective experience and hypnotic susceptibility. Relatedly, the available empirical data fail to support the hypothesis that glossolalics suffer higher levels of psychopathology than nonglossolalics."
Spanos et al then go on to detail their own research, in which they tried to teach glossolalia as a learnable skill. First, 60 subjects listened to a 60-second sample of genuine glossolalia. All subjects then tried to speak in tongues for 30 seconds. Some 20% spoke in tongues immediately without further training. The subjects were then divided into a control group and a group that received various kinds of training. Tests then showed that 70% of the trained subjects were now fluent (?) in glossolalia. Glossolalia, therefore, seems likely to be a type of learned behavior rather than a special altered state of mind.
(Spanos, Nicholas P., et al; "Glossolalia as Learned Behavior: An Experimental Demonstration," Journal of Abnormal Psychology, 95:21, 1986.)