Publication Date



Open access

Embargo Period


Degree Type


Degree Name

Doctor of Philosophy (PHD)


Biomedical Engineering (Engineering)

Date of Defense


First Committee Member

Ozcan Ozdamar

Second Committee Member

Jorge Bohorquez

Third Committee Member

Abhishek Prasad

Fourth Committee Member

Vittorio Porciatti

Fifth Committee Member

Rafael E. Delgado


Visual Evoked Potentials (VEPs) are electrophysiological signals recorded from occipital areas of the scalp in response to visual stimulation. VEPs are used for clinical monitoring of diseases and Brain-Computer Interfaces (BCIs). Depending on the stimulation rate, there are two types of VEPs, transient (low stimulation rates < 3Hz) and Steady-State (higher stimulation rates >5Hz). However, a novel VEP stimulation paradigm was developed at the University of Miami, Neurosensory Laboratory. This paradigm generates low jittered VEPs similar to Steady-State VEP (SSVEP) and called Quasi-Steady State VEPs (QSS-VEPs). QSS paradigm has proven itself in the fast acquisition of Auditory Brainstem Responses (ABRs) and Pattern Electroretinograms (PERGs). One major advantage of this paradigm is that it enables the extraction of clinically significant transient responses from QSS responses at even high rates. VEPs are extensively used in BCI research due to their advantages. SSVEPs and code-modulated VEP (c-VEPs) provided high Information Transfer Rates (ITRs), easy configuration, minimal training and encoding of multiple targets. This work introduces the QSS-VEP paradigm at high stimulation rates. Since QSS-VEP is a special type of c-VEP and similar to SSVEP, it involves the advantages of both SSVEP and c-VEP paradigms. To investigate these advantages, a dual target Brain-Computer Interface switch was developed. Two targets, left and right, were presented as horizontal bar pattern-reversal stimulation by two LED displays. In this work, the feasibility of a BCI switch by means of QSS-VEPs is provided by testing the QSS-VEP paradigm at two stimulation rates, 10 reversal per second (rps) and 50rps. Four signal types (left and right x QSS and transient) have been analyzed. 50rps has and QSS showed better performance in BCI operation compared to 10rps and transient. As another goal, the effects of stimulation rate on BCI accuracy and ITR are demonstrated by expanding the stimulation rates to four different stimulation rates, one low(10rps) , one medium (32rps) and one high (50rps) and one very high (70rps), It has been found that 32rps has the highest accuracy with a single sweep (0.5s) data, on average 43.9 bit per minute (bpm) was achieved. In the third experiment, SSVEP and QSS-VEP performances have been compared at two rates, 10rps (low) and 32rps (medium-high). For 10rps both paradigms yielded similar performances, however, for 32rps SSVEPs diminished and significantly low detection accuracies were achieved compared to 32rps QSS paradigm. In the classification for the performance evaluation, a cross-correlation coefficient based left-right classification was employed. A combined classifier based on the left and right cross-correlation coefficient resulted in better detection performance. In conclusion, the QSS-VEP paradigm is a promising alternative to SSVEP and c-VEP in a BCI system. It has the advantages of both paradigms. It has a high SNR compared to the SSVEP paradigm. It also allows the acquisition of unitary/transient VEPs at various stimulation rates, which helps with understanding the SSVEP and c-VEP generation mechanisms. This can be utilized in better designing VEP based BCI systems.


Visual Evoked Potentials; Brain-Computer Interface; Pattern-reversal VEP; Stimulation Rate; Quasi-Steady-State VEP