Master’s Scholarship Available in TÜBİTAK Project on Auditory Event Recognition

Applications are invited for a master’s level researcher position in the TUBITAK-funded project “The effect of temporal ordering and sound source in understanding the perception of sound events ”The project will offer monthly 3000 TRY for at least a year, allowance for attending inter/national conferences and an optional tuition fee waiver at Yeditepe University depending on the time of acceptance. The position is available immediately.

Applicants for a master’s level researcher position should preferably have experience with at least two of the following: programming, behavioral testing, statistical analysis, EEG analysis. The project will last up to 16 months.

To apply for this position, please email your CV, a motivation letter and contact details of two references to fyildirim[at]cse.yeditepe.edu.tr before 21 September 2020. Applicants should be enrolled in a relevant Master’s program in a Turkish university (preferably in Istanbul) by latest October 2020.

Computational Neuroscience and Visual Perception Laboratory (CNVP Lab) at Yeditepe University is located in a lively university campus, has a high collaboration profile and is affiliated with the Cognitive Science Master’s Program and the Department of Computer Engineering. The lab is equipped with a portable EEG device, a Gazepoint GP3 eye tracker and has access to a 3T MR scanner and a 32-channel EEG device located in the Yeditepe University Research Hospital.

 

The effect of temporal ordering and sound source in understanding the perception of sound events

Our performance in recognizing and identifying the source in sound perception varies according to the content of the sound. Sound waves caused by speech, music and natural sounds are identified and processed according to frequency, magnitude and phase characteristics when transmitted to the brain and sent to related regions to execute functions such as motor, attention and decision making. The perception of sound using different channels may be processed according to the semantic or physical characteristics of the sound, depending on the type and content of the information we acquire. In this study, the mechanism of event-related sounds that we are exposed to in daily life and guide our movements through passive attention processes will be studied.

We often identify the sounds we hear at certain noise ratios based on our previous experience. However, when some characteristics of the sound are obscured, it may be impossible to identify. In this study, we will examine how the perception process of sound events with cause-effect relationship is processed depending on the temporal sequence and type of event. In cases where the temporal order of the sound is distorted at different rates, local or global distortion may occur by maintaining consecutive information. The first part of the project will be about examining the effect of these distortion conditions on the performance and determining the temporal parameters necessary for the identification of sound events. These parameters will help understanding which of the physical characteristics of sound or the time windows of the events are more important for performance.

In the second part of the study, the relationship between the identification of sounds that contain predictable sound events (e.g. car braking and crashing, rolling coin toppling, etc.) and predictability of the second phase of sound event will be examined using event related potentials (ERPs). EEG data will be evaluated using frequency and correlational analysis methods. The effect of timewise relations (e.g. neighborhood, ranking of general units, fine-sorting) will be investigated with regards to identification performance of the sounds.

%d bloggers like this: