scholarly journals “Surprise” Neurons in the Supplementary Eye Field of the Cerebral Cortex and Exploration-exploitation Trade-off

2016 ◽  
Vol 23 (1) ◽  
pp. 41-45
Author(s):  
Kazuhiro Sakamoto ◽  
Norihiko Kawaguchi ◽  
Hajime Mushiake
2015 ◽  
Vol 113 (3) ◽  
pp. 1001-1014 ◽  
Author(s):  
Norihiko Kawaguchi ◽  
Kazuhiro Sakamoto ◽  
Naohiro Saito ◽  
Yoshito Furusawa ◽  
Jun Tanji ◽  
...  

Visual search is coordinated adaptively by monitoring and predicting the environment. The supplementary eye field (SEF) plays a role in oculomotor control and outcome evaluation. However, it is not clear whether the SEF is involved in adjusting behavioral modes based on preceding feedback. We hypothesized that the SEF drives exploration-exploitation transitions by generating “surprise signals” or rectified prediction errors, which reflect differences between predicted and actual outcomes. To test this hypothesis, we introduced an oculomotor two-target search task in which monkeys were required to find two valid targets among four identical stimuli. After they detected the valid targets, they exploited their knowledge of target locations to obtain a reward by choosing the two valid targets alternately. Behavioral analysis revealed two distinct types of oculomotor search patterns: exploration and exploitation. We found that two types of SEF neurons represented the surprise signals. The error-surprise neurons showed enhanced activity when the monkey received the first error feedback after the target pair change, and this activity was followed by an exploratory oculomotor search pattern. The correct-surprise neurons showed enhanced activity when the monkey received the first correct feedback after an error trial, and this increased activity was followed by an exploitative, fixed-type search pattern. Our findings suggest that error-surprise neurons are involved in the transition from exploitation to exploration and that correct-surprise neurons are involved in the transition from exploration to exploitation.


1996 ◽  
Vol 75 (5) ◽  
pp. 2187-2191 ◽  
Author(s):  
H. Mushiake ◽  
N. Fujii ◽  
J. Tanji

1. We studied neuronal activity in the supplementary eye field (SEF) and frontal eye field (FEF) of a monkey during performance of a conditional motor task that required capturing of a target either with a saccadic eye movement (the saccade-only condition) or with an eye-hand reach (the saccade-and-reach condition), according to visual instructions. 2. Among 106 SEF neurons that showed presaccadic activity, more than one-half of them (54%) were active preferentially under the saccade-only condition (n = 12) or under the saccade-and-reach condition (n = 45), while the remaining 49 neurons were equally active in both conditions. 3. By contrast, most (97%) of the 109 neurons in the FEF exhibited approximately equal activity in relation to saccades under the two conditions. 4. The present results suggest the possibility that SEF neurons, at least in part, are involved in signaling whether the motor task is oculomotor or combined eye-arm movements, whereas FEF neurons are mostly related to oculomotor control.


1987 ◽  
Vol 57 (1) ◽  
pp. 179-200 ◽  
Author(s):  
J. Schlag ◽  
M. Schlag-Rey

Electrical microstimulation and unit recording were performed in dorsomedial frontal cortex of four alert monkeys to identify an oculomotor area whose existence had been postulated rostral to the supplementary motor area. Contraversive saccades were evoked from 129 sites by stimulation. Threshold currents were lower than 20 microA in half the tests. Response latencies were usually longer than 50 ms (minimum: 30 ms). Eye movements were occasionally accompanied by blinks, ear, or neck movements. The cortical area yielding these movements was at the superior edge of the frontal lobe just rostral to the region from which limb movements could be elicited. Depending on the site of stimulation, saccades varied between two extremes: from having rather uniform direction and size, to converging toward a goal defined in space. The transition between these extremes was gradual with no evidence that these two types were fundamentally different. From surface to depth of cortex, direction and amplitude of evoked saccades were similar or changed progressively. No clear systematization was found depending on location along rostrocaudal or mediolateral axes of the cortex. The dorsomedial oculomotor area mapped was approximately 7 mm long and 6 mm wide. Combined eye and head movements were elicited from one of ten sites stimulated when the head was unrestrained. In the other nine cases, saccades were not accompanied by head rotation, even when higher currents or longer stimulus trains were applied. Presaccadic unit activity was recorded from 62 cells. Each of these cells had a preferred direction that corresponded to the direction of the movement evoked by local microstimulation. Presaccadic activity occurred with self-initiated as well as visually triggered saccades. It often led self-initiated saccades by more than 300 ms. Recordings made with the head free showed that the firing could not be interpreted as due to attempted head movements. Many dorsomedial cortical neurons responded to photic stimuli, either phasically or tonically. Sustained responses (activation or inhibition) were observed during target fixation. Twenty-one presaccadic units showed tonic changes of activity with fixation. Justification is given for considering the cortical area studied as a supplementary eye field. It shares many common properties with the arcuate frontal eye field. Differences noted in this study include: longer latency of response to electrical stimulation, possibility to evoke saccades converging apparently toward a goal, and long-lead unit activity with spontaneous saccades.


2011 ◽  
Vol 11 (11) ◽  
pp. 529-529
Author(s):  
S. Heinen ◽  
S.-n. Yang ◽  
J. Ford

Author(s):  
Julian Berk ◽  
Sunil Gupta ◽  
Santu Rana ◽  
Svetha Venkatesh

In order to improve the performance of Bayesian optimisation, we develop a modified Gaussian process upper confidence bound (GP-UCB) acquisition function. This is done by sampling the exploration-exploitation trade-off parameter from a distribution. We prove that this allows the expected trade-off parameter to be altered to better suit the problem without compromising a bound on the function's Bayesian regret. We also provide results showing that our method achieves better performance than GP-UCB in a range of real-world and synthetic problems.


2009 ◽  
Author(s):  
Antoine Soubeyran ◽  
Raphael Soubeyran ◽  
Ngo Van Long

2019 ◽  
Vol 19 (10) ◽  
pp. 306c
Author(s):  
Steven P Errington ◽  
Amirsaman Sajad ◽  
Jeffrey D Schall

Sign in / Sign up

Export Citation Format

Share Document