Marco Congedo

Date of Award


Degree Type


Degree Name

Doctor of Philosophy



Major Professor

Joel F. Lubar


A major limitation of the current neurofeedback paradigm is the limited information provided by a single or a small number of electrodes placed on the scalp. A considerable improvement of the neurofeedback efficacy and specificity could be obtained feeding back brain activity of delimited structures. While traditional EEG information reflects the superposition of the electrical activity of a large number of neurons, by means of inverse solutions such as the Low-Resolution Electromagnetic Tomography (LORETA) spatially delimited brain activity can be evaluated in neocortical tissue. In this Dissertation we implement LORETA neurofeedback, we introduce a new feedback function ( 1 ) sensitive to dynamic change over time, and we clarify several issues related to the learning process observable with neurofeedback. The reported set of three experiments is the first attempt I am aware of to prove learning of brain current density activity. Three individuals were trained to improve brain activation (suppress low Alpha (8-10 Hz) and enhance low Beta (16 -20 Hz) current density) in the anterior cingulate gyros cognitive division (ACcd). Participants took part of six experimental sessions, each lasting approximately 30 minutes. Randomization-Permutation ANCOVA tests were conducted on recordings of the neurofeedback training. In addition a randomized trial was performed at the end of the treatment. During eight two-minutes periods (trials) participants were asked to try to obtain as many rewards as they could ( 4 "1 " trials) or as few rewards as they could ( 4 "0" trials). The order of trials was decided at random. The hypothesis under testing was that participants acquired volitional control over their brain activity so to be able to obtain more rewards during the plus condition as compared to the minus condition. We found evidence of volitional control for two subjects (p=0.043 and p=O.l) and no evidence of volitional control for one of them (p=0.27 1). The combination of the three p-values provided an overall probability value for this experiment of 0.012 with the additive method and 0.035 with the multiplicative method. These results strongly support the hypothesis of volitional control across the experimental group. Trends of the Beta/ Alpha power ratio in the ACcd were in the expected direction for all the three subjects, however the combined p-values did not reach significance. With as few as six training sessions, typically insufficient to produce any form of learning with scalp neurofeedback, the experiment showed overall signs of volitional control of the electrical activity of the ACcd. Possible applications of the technique are important and include the treatment of epileptic foci, the treatment of specific brain regions damaged as a consequence of traumatic brain injury, and in general of any specific cortical electrical activity.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."