(from the chapter) In this chapter, we identify two different subtypes of number-color synesthesi... more (from the chapter) In this chapter, we identify two different subtypes of number-color synesthesia and propose that they are caused by hyperconnectivity between color and number areas at different stages in processing. "Lower synesthetes" may have cross-wiring (or cross-activation) within the fusiform gyrus, whereas "higher synesthetes" may have cross-activation in the angular gyrus. We then discuss the implications of the study of synesthesia for understanding other aspects of the human mind, such as creativity and metaphor. Finally, we discuss the implications of synesthesia for the philosophical problem of qualia. (PsycINFO Database Record (c) 2010 APA, all rights reserved) (chapter)
Motor learning has recently been shown to involve both error-based feedback and reward-based feed... more Motor learning has recently been shown to involve both error-based feedback and reward-based feedback. It has been suggested that error information is critical for an incremental descent process that converges towards the appropriate motor command while reward information is critical for motor command consolidation. In terms of neural systems, the cerebellum has been hypothesized to play a critical role in the generation and utilization of errors signals for learning, whereas the motor cortex has been hypothesized to be a site of consolidation. To explore these hypotheses, we applied transcranial magnetic stimulation (TMS) over motor cortex during visuomotor adaptation. We compared two conditions; applying TMS so as to perturb neural activity on either rewarded or unrewarded trials. If motor cortex is primarily involved in consolidation, we predicted that TMS on rewarded trials would be disruptive. In contrast, TMS on unrewarded trials should have little effect on performance.
Participants made reaching movements to visual targets without vision of the limb. If the movement intersected the target, the target turned green and a pleasant “ding” was played. If the movement missed the target, the target turned red and an aversive “buzz” was played. A financial bonus system was used to increase motivation given the challenges of this categorical form of feedback. During a series of 441 reaches, a visuomotor rotation was gradually introduced between the “virtual” cursor location and the hand location, reaching a maximum of 8 degrees displacement (Izawa and Shadmehr, 2011). At the end of this adaptation phase, participants completed another 49 trials without any feedback. This washout phase allowed us to measure the temporal decay of motor learning. Participants were assigned to one of three groups. In the Success group, a TMS pulse was delivered on trials in which the reach intersected the target. In the Failure group, a TMS pulse was delivered on trials in which the reach missed the target. In the Control group, no TMS pulses were delivered.
The Failure and Control groups showed similar levels of learning during the adaptation phase, as well as similar decay during the washout phase. In contrast, the Success group did not learn the rotation and, as such, showed no aftereffect. While additional controls are required to rule out non-specific hypotheses (e.g., TMS is aversive, independent of the targeted location), the results suggest that disruption over motor cortex at the time of reward impairs motor adaptation. This would be consistent with the hypothesis that during normal learning, reward-signals following successful movements facilitate consolidation processes.
Some otherwise normal individuals experience sensations in one sensory modality when a second mod... more Some otherwise normal individuals experience sensations in one sensory modality when a second modality is stimulated. For example, a synesthete may see any given number as always tinged a certain color (e.g. 5, may be green and 6, may be red). We have previously shown that synesthetically induced colors can influence perceptual grouping, can lead to pop-out and are not affected by crowding (a form of blindsight). This suggests that synesthesia is a sensory phenomenon caused by cross activation of brain maps corresponding to color (V4 or V8) and graphemes, which lie adjacent to each other in the fusiform gyrus (Ramachandran and Hubbard, SFN 2000, 2001 VSS 2001). We now present evidence that top-down influences can also affect synesthetic colors. We find: 1) When presented with hierarchical figures (e.g., a large 5 composed of small 2's) the colors depend on where a synesthete focuses his attention (the exact experience differs depending on whether letters or numbers are used) 2) For synesthetes who experience letters in color, unitization (for example, missing the fs in of) can affect whether or not a synesthete experiences colors for letters and 3) Most synesthetes report that imagining a grapheme evokes the corresponding color more strongly than actually seeing it. We tested two synesthetic subjects (JC and ER) on a variant of the Perky experiment, in which mental imagery interferes with visual perception. Subjects made a 2AFC identification of colored letters or numbers while imagining either consistent or inconsistent graphemes. When their performance with and without imagery were compared, synesthetes showed a much greater increase in threshold for detecting the presence of a grapheme when they were imagining another grapheme then did normals (a larger Perky effect). We therefore conclude that although synesthetic colors arise early in perception, they can also be influenced by top-down influences.
Although synesthesia (“seeing sounds” or seeing letters and numbers tinged with specific colors) ... more Although synesthesia (“seeing sounds” or seeing letters and numbers tinged with specific colors) has been known for over 100 years, only recently has progress been made towards understanding its underlying mechanisms. We have previously demonstrated that synesthesia is a perceptual phenomenon, and suggest that, in some synesthetes, it may be caused by cross-wiring between V4 and the number grapheme area, both of which lie in the fusiform gyrus (Ramachandran & Hubbard, SFN 2000; 2001, VSS 2001). Roman numbers and subitizable clusters of dots did not evoke colors, suggesting that the visual grapheme is critical for the experience of colors. To explore this further, we presented graphemes in unusual fonts, in lower case, or made from unusual forms (i.e. a drawing of three cats to make up the form of an “A”). Synesthete JC reported that he still experienced the appropriate synesthetic colors, but lighter and “shinier.” These results further show that the visual grapheme is critical for the experience of colors. However, we have also identified a second group of “higher synesthetes” in whom the colors are driven by the numerical concept (e.g., by Roman numerals). For these synesthetes, the experience of colors does not change depending on visual properties such as font. We propose that in this group of synesthetes, cross-wiring may occur near the angular gyrus, where higher color areas and more abstract number areas lie. Intriguingly, in some of these synesthetes, even days of the week and months of the year are colored. We therefore suggest that the angular gyrus may be involved in representing abstract sequences or ordinality, and this would explain why numbers, days of the week and month of the year are colored in these synesthetes. This heterogeneity has often led researchers to avoid studying synesthesia. Our results suggest that there may be an underlying logic to the phenomena, and may help to guide further research into the neural basis of synesthesia.
An associative learning task was employed in which, in separate blocks, reinforcement was limited... more An associative learning task was employed in which, in separate blocks, reinforcement was limited to either correct or incorrect trials. Given the role of dopamine in reinforcement learning, we examined the effects of a dopamine (DA) agonist on performance and obtained genetic information to divide the participants based on two SNPs related to intrinsic DA levels, COMT and DRD2.
Participants learned to associate 16 abstract images with one of two response categories. Each trial had a 50% chance of being eligible for informative feedback. If ineligible, the response was followed by the feedback message “Unknown”. In the positive feedback condition, the feedback message “Correct” was displayed if the trial was eligible for feedback and the response was correct; otherwise the message “Unknown” was shown. Similarly, the message “Incorrect” was presented on eligible trials in the negative feedback condition when the answer was incorrect. In this manner, we separated positive and negative feedback-based learning. Participants were tested in two sessions in a double blind study. In one session participants ingested 1.25 mg of Bromocriptine (BR), a DA agonist; in the other session, they were given a placebo (PLAC).
On placebo, performance was more accurate in the positive condition (90%) than in the negative condition (83%). Overall, there was a Drug x Reinforcement Type interaction: BR resulted in a drop in accuracy for the positive condition and an increase in accuracy for negative condition. The reduction in performance from BR during positive reinforcement was similar for all alleles of COMT. In contrast, the benefit from BR during negative reinforcement was restricted to those individuals with higher levels of intrinsic DA (Val/Met or Met/MET alleles). For DRD2, the interaction of BR and reinforcement type was most pronounced in individuals with lower levels of intrinsic DA (A/A or A/G). These results suggest that individuals with higher DA levels in the frontal lobe may benefit in learning from negative feedback when DA signaling is increased, while individuals with lower striatal levels of DA may benefit in learning from both types of feedback with increased DA signaling.
F87
EFFECTS OF THETA BURST STIMULATION OF MIDDLE FRONTAL GYRUS AND INFERIOR PARIETAL LOBULE ON L... more F87
EFFECTS OF THETA BURST STIMULATION OF MIDDLE FRONTAL GYRUS AND INFERIOR PARIETAL LOBULE ON LOAD AND FILTERING PROCESSES IN SPATIAL WORKING MEMORY
Peter A Butcher1, Richard B Ivry1;
1Department of Psychology, UC Berkeley
The exclusion of task-irrelevant information from working memory has been associated with activation in the middle frontal gyrus (MFG) and basal ganglia (McNab, 2007). In contrast, activation in right inferior parietal lobule (IPL) is associated with WM load. To further explore these hypotheses, we disrupted MFG and IPL in the right hemisphere with offline theta burst stimulation (TBS). On each trial, subjects viewed a display of 14 squares arranged along a virtual circle about a fixation point. Three different conditions were verbally cued for each trial. Low Load: Four red circles indicated possible target location. High Load: Memory load was increased to six items, 4 red and 2 yellow circles. Filter: 4 red and 2 yellow circles were presented but yellow circles were ignored. After a delay, a probe appeared and participants indicated if it was one of the target locations. While TBS to rMFG and rIPL produced minimal change in load cost (Low-High) or filter cost (Low-Filter), there was a change in the correlation pattern of the two measures across individuals. The measures were strongly correlated following control stimulation. Following TBS to either area, these correlations were reduced for contralateral targets, but remained unchanged for ipsilateral targets. Thus, while TBS of either region affected performance for contralateral targets, we did not observe dissociable contributions of MFG and IPL to spatial WM.
Working memory (WM) requires selecting and maintaining task-relevant information, and the exclu... more Working memory (WM) requires selecting and maintaining task-relevant information, and the exclusion of task-irrelevant information. A recent fMRI study (McNab, 2008) indicated that middle frontal gyrus and the basal ganglia are more strongly activated when stimulus displays include information to be ignored and maintained, as compared to when they just include information to be maintained. Given that these frontal-striatal regions are strongly modulated by the brainstem dopaminergic system, we examined the role of dopamine (DA) in this filtering process.
Subjects were tested on two sessions in a double blind behavioral study. In one session, subjects were given 1.25 mg of bromocriptine, a DA agonist, and in the other session, a placebo (PLAC). Using the design of McNab (2008), subjects performed a spatial WM task. On each trial, subjects viewed a display composed of 14 squares arranged along a virtual circle about a fixation point. Three different conditions were verbally cued at the start of each trial. In the "Low Load" condition, subjects were instructed to remember the location of 4 red circles. In the "High Load" condition, the memory load was increased to six items, 4 red and 2 yellow circles. In the "Filter" condition, 4 red and 2 yellow circles were presented but subjects were instructed to only remember the locations marked with red squares. The memory set was presented for 200 ms, followed by a 2 s delay period, followed by a probe, either within one of the to-be-remembered locations, or at a non-target location. Subjects indicated whether the probed location was at one of the target locations.
Subjects were divided into two groups, based on their performance on their spatial WM span. The High Span group was more accurate than the Low Span group after placebo (78% vs. 72%). Bromocriptine had minimal effects on accuracies, but larger RT effects were observed. Across all three WM conditions, the High Span group showed RT increases on bromocriptine as compared to placebo. In contrast, the Low Span group showed RT decreases on Bromocriptine. Both groups showed attenuated load and filtering costs after bromocriptine administration. Indeed neither group showed a RT increase on the Filter trials compared to the Low Load trials, suggesting that the dopamine either taxed performance on the Low Load trials or led to near-perfect filtering. Our results suggest two conclusions regarding DA and WM function. First, consistent with previous studies demonstrating an inverted U-shaped function relating dopamine and WM performance, high and low span individuals were affected in opposite directions by DA augmentation. Second, DA may reduce costs associated with increased demands on WM.
A network spanning prefrontal cortex and the basal ganglia (BG) have been implicated in an fMRI... more A network spanning prefrontal cortex and the basal ganglia (BG) have been implicated in an fMRI study in filtering information that gains access to visual working memory (WM). Specifically, bilateral middle frontal gyrus and the left BG were more active on trials in which distractors were present with trials compared to trials in which no distractors were presented. The lateralized activity in the BG suggests that filtering may be better when distractors are presented in the right visual field (RVF). We conducted a behavioral study to test this prediction. Participants were shown a display composed of 14 squares arranged in a circle centered at the fixation point. On each trial, red and yellow circles appeared in a subset of the squares. Prior to the memory display, one of two instruction cues had been presented. In the “red only” condition, participants had to remember only the locations that had red circles inside of them; they were to ignore the yellow circles (if present). In the “all” condition, participants were instructed to remember the locations of both the red and yellow circles. After a two-second delay, a question mark appeared in one location and participants indicated with a 2-AFC judgment whether the probed location had contained one of the circles in the working memory set. The memory display was presented for only 200 ms to minimize the utility of eye movements, and thus, keep the stimuli lateralized. Participants were consistently better on the memory test when locations in the left visual field were probed (80% vs. 78% overall accuracy). However, the visual field advantage was similar on filtering (4%) and non-filtering (4%) trials. While the results indicate there may be hemispheric (or visual field) differences in WM performance, the results failed to demonstrate any asymmetries in filtering effects. Further investigation is necessary to identify components of working memory that may account for the LVF advantage.
(from the chapter) In this chapter, we identify two different subtypes of number-color synesthesi... more (from the chapter) In this chapter, we identify two different subtypes of number-color synesthesia and propose that they are caused by hyperconnectivity between color and number areas at different stages in processing. "Lower synesthetes" may have cross-wiring (or cross-activation) within the fusiform gyrus, whereas "higher synesthetes" may have cross-activation in the angular gyrus. We then discuss the implications of the study of synesthesia for understanding other aspects of the human mind, such as creativity and metaphor. Finally, we discuss the implications of synesthesia for the philosophical problem of qualia. (PsycINFO Database Record (c) 2010 APA, all rights reserved) (chapter)
Motor learning has recently been shown to involve both error-based feedback and reward-based feed... more Motor learning has recently been shown to involve both error-based feedback and reward-based feedback. It has been suggested that error information is critical for an incremental descent process that converges towards the appropriate motor command while reward information is critical for motor command consolidation. In terms of neural systems, the cerebellum has been hypothesized to play a critical role in the generation and utilization of errors signals for learning, whereas the motor cortex has been hypothesized to be a site of consolidation. To explore these hypotheses, we applied transcranial magnetic stimulation (TMS) over motor cortex during visuomotor adaptation. We compared two conditions; applying TMS so as to perturb neural activity on either rewarded or unrewarded trials. If motor cortex is primarily involved in consolidation, we predicted that TMS on rewarded trials would be disruptive. In contrast, TMS on unrewarded trials should have little effect on performance.
Participants made reaching movements to visual targets without vision of the limb. If the movement intersected the target, the target turned green and a pleasant “ding” was played. If the movement missed the target, the target turned red and an aversive “buzz” was played. A financial bonus system was used to increase motivation given the challenges of this categorical form of feedback. During a series of 441 reaches, a visuomotor rotation was gradually introduced between the “virtual” cursor location and the hand location, reaching a maximum of 8 degrees displacement (Izawa and Shadmehr, 2011). At the end of this adaptation phase, participants completed another 49 trials without any feedback. This washout phase allowed us to measure the temporal decay of motor learning. Participants were assigned to one of three groups. In the Success group, a TMS pulse was delivered on trials in which the reach intersected the target. In the Failure group, a TMS pulse was delivered on trials in which the reach missed the target. In the Control group, no TMS pulses were delivered.
The Failure and Control groups showed similar levels of learning during the adaptation phase, as well as similar decay during the washout phase. In contrast, the Success group did not learn the rotation and, as such, showed no aftereffect. While additional controls are required to rule out non-specific hypotheses (e.g., TMS is aversive, independent of the targeted location), the results suggest that disruption over motor cortex at the time of reward impairs motor adaptation. This would be consistent with the hypothesis that during normal learning, reward-signals following successful movements facilitate consolidation processes.
Some otherwise normal individuals experience sensations in one sensory modality when a second mod... more Some otherwise normal individuals experience sensations in one sensory modality when a second modality is stimulated. For example, a synesthete may see any given number as always tinged a certain color (e.g. 5, may be green and 6, may be red). We have previously shown that synesthetically induced colors can influence perceptual grouping, can lead to pop-out and are not affected by crowding (a form of blindsight). This suggests that synesthesia is a sensory phenomenon caused by cross activation of brain maps corresponding to color (V4 or V8) and graphemes, which lie adjacent to each other in the fusiform gyrus (Ramachandran and Hubbard, SFN 2000, 2001 VSS 2001). We now present evidence that top-down influences can also affect synesthetic colors. We find: 1) When presented with hierarchical figures (e.g., a large 5 composed of small 2's) the colors depend on where a synesthete focuses his attention (the exact experience differs depending on whether letters or numbers are used) 2) For synesthetes who experience letters in color, unitization (for example, missing the fs in of) can affect whether or not a synesthete experiences colors for letters and 3) Most synesthetes report that imagining a grapheme evokes the corresponding color more strongly than actually seeing it. We tested two synesthetic subjects (JC and ER) on a variant of the Perky experiment, in which mental imagery interferes with visual perception. Subjects made a 2AFC identification of colored letters or numbers while imagining either consistent or inconsistent graphemes. When their performance with and without imagery were compared, synesthetes showed a much greater increase in threshold for detecting the presence of a grapheme when they were imagining another grapheme then did normals (a larger Perky effect). We therefore conclude that although synesthetic colors arise early in perception, they can also be influenced by top-down influences.
Although synesthesia (“seeing sounds” or seeing letters and numbers tinged with specific colors) ... more Although synesthesia (“seeing sounds” or seeing letters and numbers tinged with specific colors) has been known for over 100 years, only recently has progress been made towards understanding its underlying mechanisms. We have previously demonstrated that synesthesia is a perceptual phenomenon, and suggest that, in some synesthetes, it may be caused by cross-wiring between V4 and the number grapheme area, both of which lie in the fusiform gyrus (Ramachandran & Hubbard, SFN 2000; 2001, VSS 2001). Roman numbers and subitizable clusters of dots did not evoke colors, suggesting that the visual grapheme is critical for the experience of colors. To explore this further, we presented graphemes in unusual fonts, in lower case, or made from unusual forms (i.e. a drawing of three cats to make up the form of an “A”). Synesthete JC reported that he still experienced the appropriate synesthetic colors, but lighter and “shinier.” These results further show that the visual grapheme is critical for the experience of colors. However, we have also identified a second group of “higher synesthetes” in whom the colors are driven by the numerical concept (e.g., by Roman numerals). For these synesthetes, the experience of colors does not change depending on visual properties such as font. We propose that in this group of synesthetes, cross-wiring may occur near the angular gyrus, where higher color areas and more abstract number areas lie. Intriguingly, in some of these synesthetes, even days of the week and months of the year are colored. We therefore suggest that the angular gyrus may be involved in representing abstract sequences or ordinality, and this would explain why numbers, days of the week and month of the year are colored in these synesthetes. This heterogeneity has often led researchers to avoid studying synesthesia. Our results suggest that there may be an underlying logic to the phenomena, and may help to guide further research into the neural basis of synesthesia.
An associative learning task was employed in which, in separate blocks, reinforcement was limited... more An associative learning task was employed in which, in separate blocks, reinforcement was limited to either correct or incorrect trials. Given the role of dopamine in reinforcement learning, we examined the effects of a dopamine (DA) agonist on performance and obtained genetic information to divide the participants based on two SNPs related to intrinsic DA levels, COMT and DRD2.
Participants learned to associate 16 abstract images with one of two response categories. Each trial had a 50% chance of being eligible for informative feedback. If ineligible, the response was followed by the feedback message “Unknown”. In the positive feedback condition, the feedback message “Correct” was displayed if the trial was eligible for feedback and the response was correct; otherwise the message “Unknown” was shown. Similarly, the message “Incorrect” was presented on eligible trials in the negative feedback condition when the answer was incorrect. In this manner, we separated positive and negative feedback-based learning. Participants were tested in two sessions in a double blind study. In one session participants ingested 1.25 mg of Bromocriptine (BR), a DA agonist; in the other session, they were given a placebo (PLAC).
On placebo, performance was more accurate in the positive condition (90%) than in the negative condition (83%). Overall, there was a Drug x Reinforcement Type interaction: BR resulted in a drop in accuracy for the positive condition and an increase in accuracy for negative condition. The reduction in performance from BR during positive reinforcement was similar for all alleles of COMT. In contrast, the benefit from BR during negative reinforcement was restricted to those individuals with higher levels of intrinsic DA (Val/Met or Met/MET alleles). For DRD2, the interaction of BR and reinforcement type was most pronounced in individuals with lower levels of intrinsic DA (A/A or A/G). These results suggest that individuals with higher DA levels in the frontal lobe may benefit in learning from negative feedback when DA signaling is increased, while individuals with lower striatal levels of DA may benefit in learning from both types of feedback with increased DA signaling.
F87
EFFECTS OF THETA BURST STIMULATION OF MIDDLE FRONTAL GYRUS AND INFERIOR PARIETAL LOBULE ON L... more F87
EFFECTS OF THETA BURST STIMULATION OF MIDDLE FRONTAL GYRUS AND INFERIOR PARIETAL LOBULE ON LOAD AND FILTERING PROCESSES IN SPATIAL WORKING MEMORY
Peter A Butcher1, Richard B Ivry1;
1Department of Psychology, UC Berkeley
The exclusion of task-irrelevant information from working memory has been associated with activation in the middle frontal gyrus (MFG) and basal ganglia (McNab, 2007). In contrast, activation in right inferior parietal lobule (IPL) is associated with WM load. To further explore these hypotheses, we disrupted MFG and IPL in the right hemisphere with offline theta burst stimulation (TBS). On each trial, subjects viewed a display of 14 squares arranged along a virtual circle about a fixation point. Three different conditions were verbally cued for each trial. Low Load: Four red circles indicated possible target location. High Load: Memory load was increased to six items, 4 red and 2 yellow circles. Filter: 4 red and 2 yellow circles were presented but yellow circles were ignored. After a delay, a probe appeared and participants indicated if it was one of the target locations. While TBS to rMFG and rIPL produced minimal change in load cost (Low-High) or filter cost (Low-Filter), there was a change in the correlation pattern of the two measures across individuals. The measures were strongly correlated following control stimulation. Following TBS to either area, these correlations were reduced for contralateral targets, but remained unchanged for ipsilateral targets. Thus, while TBS of either region affected performance for contralateral targets, we did not observe dissociable contributions of MFG and IPL to spatial WM.
Working memory (WM) requires selecting and maintaining task-relevant information, and the exclu... more Working memory (WM) requires selecting and maintaining task-relevant information, and the exclusion of task-irrelevant information. A recent fMRI study (McNab, 2008) indicated that middle frontal gyrus and the basal ganglia are more strongly activated when stimulus displays include information to be ignored and maintained, as compared to when they just include information to be maintained. Given that these frontal-striatal regions are strongly modulated by the brainstem dopaminergic system, we examined the role of dopamine (DA) in this filtering process.
Subjects were tested on two sessions in a double blind behavioral study. In one session, subjects were given 1.25 mg of bromocriptine, a DA agonist, and in the other session, a placebo (PLAC). Using the design of McNab (2008), subjects performed a spatial WM task. On each trial, subjects viewed a display composed of 14 squares arranged along a virtual circle about a fixation point. Three different conditions were verbally cued at the start of each trial. In the "Low Load" condition, subjects were instructed to remember the location of 4 red circles. In the "High Load" condition, the memory load was increased to six items, 4 red and 2 yellow circles. In the "Filter" condition, 4 red and 2 yellow circles were presented but subjects were instructed to only remember the locations marked with red squares. The memory set was presented for 200 ms, followed by a 2 s delay period, followed by a probe, either within one of the to-be-remembered locations, or at a non-target location. Subjects indicated whether the probed location was at one of the target locations.
Subjects were divided into two groups, based on their performance on their spatial WM span. The High Span group was more accurate than the Low Span group after placebo (78% vs. 72%). Bromocriptine had minimal effects on accuracies, but larger RT effects were observed. Across all three WM conditions, the High Span group showed RT increases on bromocriptine as compared to placebo. In contrast, the Low Span group showed RT decreases on Bromocriptine. Both groups showed attenuated load and filtering costs after bromocriptine administration. Indeed neither group showed a RT increase on the Filter trials compared to the Low Load trials, suggesting that the dopamine either taxed performance on the Low Load trials or led to near-perfect filtering. Our results suggest two conclusions regarding DA and WM function. First, consistent with previous studies demonstrating an inverted U-shaped function relating dopamine and WM performance, high and low span individuals were affected in opposite directions by DA augmentation. Second, DA may reduce costs associated with increased demands on WM.
A network spanning prefrontal cortex and the basal ganglia (BG) have been implicated in an fMRI... more A network spanning prefrontal cortex and the basal ganglia (BG) have been implicated in an fMRI study in filtering information that gains access to visual working memory (WM). Specifically, bilateral middle frontal gyrus and the left BG were more active on trials in which distractors were present with trials compared to trials in which no distractors were presented. The lateralized activity in the BG suggests that filtering may be better when distractors are presented in the right visual field (RVF). We conducted a behavioral study to test this prediction. Participants were shown a display composed of 14 squares arranged in a circle centered at the fixation point. On each trial, red and yellow circles appeared in a subset of the squares. Prior to the memory display, one of two instruction cues had been presented. In the “red only” condition, participants had to remember only the locations that had red circles inside of them; they were to ignore the yellow circles (if present). In the “all” condition, participants were instructed to remember the locations of both the red and yellow circles. After a two-second delay, a question mark appeared in one location and participants indicated with a 2-AFC judgment whether the probed location had contained one of the circles in the working memory set. The memory display was presented for only 200 ms to minimize the utility of eye movements, and thus, keep the stimuli lateralized. Participants were consistently better on the memory test when locations in the left visual field were probed (80% vs. 78% overall accuracy). However, the visual field advantage was similar on filtering (4%) and non-filtering (4%) trials. While the results indicate there may be hemispheric (or visual field) differences in WM performance, the results failed to demonstrate any asymmetries in filtering effects. Further investigation is necessary to identify components of working memory that may account for the LVF advantage.
Uploads
Participants made reaching movements to visual targets without vision of the limb. If the movement intersected the target, the target turned green and a pleasant “ding” was played. If the movement missed the target, the target turned red and an aversive “buzz” was played. A financial bonus system was used to increase motivation given the challenges of this categorical form of feedback. During a series of 441 reaches, a visuomotor rotation was gradually introduced between the “virtual” cursor location and the hand location, reaching a maximum of 8 degrees displacement (Izawa and Shadmehr, 2011). At the end of this adaptation phase, participants completed another 49 trials without any feedback. This washout phase allowed us to measure the temporal decay of motor learning. Participants were assigned to one of three groups. In the Success group, a TMS pulse was delivered on trials in which the reach intersected the target. In the Failure group, a TMS pulse was delivered on trials in which the reach missed the target. In the Control group, no TMS pulses were delivered.
The Failure and Control groups showed similar levels of learning during the adaptation phase, as well as similar decay during the washout phase. In contrast, the Success group did not learn the rotation and, as such, showed no aftereffect. While additional controls are required to rule out non-specific hypotheses (e.g., TMS is aversive, independent of the targeted location), the results suggest that disruption over motor cortex at the time of reward impairs motor adaptation. This would be consistent with the hypothesis that during normal learning, reward-signals following successful movements facilitate consolidation processes.
Participants learned to associate 16 abstract images with one of two response categories. Each trial had a 50% chance of being eligible for informative feedback. If ineligible, the response was followed by the feedback message “Unknown”. In the positive feedback condition, the feedback message “Correct” was displayed if the trial was eligible for feedback and the response was correct; otherwise the message “Unknown” was shown. Similarly, the message “Incorrect” was presented on eligible trials in the negative feedback condition when the answer was incorrect. In this manner, we separated positive and negative feedback-based learning. Participants were tested in two sessions in a double blind study. In one session participants ingested 1.25 mg of Bromocriptine (BR), a DA agonist; in the other session, they were given a placebo (PLAC).
On placebo, performance was more accurate in the positive condition (90%) than in the negative condition (83%). Overall, there was a Drug x Reinforcement Type interaction: BR resulted in a drop in accuracy for the positive condition and an increase in accuracy for negative condition. The reduction in performance from BR during positive reinforcement was similar for all alleles of COMT. In contrast, the benefit from BR during negative reinforcement was restricted to those individuals with higher levels of intrinsic DA (Val/Met or Met/MET alleles). For DRD2, the interaction of BR and reinforcement type was most pronounced in individuals with lower levels of intrinsic DA (A/A or A/G). These results suggest that individuals with higher DA levels in the frontal lobe may benefit in learning from negative feedback when DA signaling is increased, while individuals with lower striatal levels of DA may benefit in learning from both types of feedback with increased DA signaling.
EFFECTS OF THETA BURST STIMULATION OF MIDDLE FRONTAL GYRUS AND INFERIOR PARIETAL LOBULE ON LOAD AND FILTERING PROCESSES IN SPATIAL WORKING MEMORY
Peter A Butcher1, Richard B Ivry1;
1Department of Psychology, UC Berkeley
The exclusion of task-irrelevant information from working memory has been associated with activation in the middle frontal gyrus (MFG) and basal ganglia (McNab, 2007). In contrast, activation in right inferior parietal lobule (IPL) is associated with WM load. To further explore these hypotheses, we disrupted MFG and IPL in the right hemisphere with offline theta burst stimulation (TBS). On each trial, subjects viewed a display of 14 squares arranged along a virtual circle about a fixation point. Three different conditions were verbally cued for each trial. Low Load: Four red circles indicated possible target location. High Load: Memory load was increased to six items, 4 red and 2 yellow circles. Filter: 4 red and 2 yellow circles were presented but yellow circles were ignored. After a delay, a probe appeared and participants indicated if it was one of the target locations. While TBS to rMFG and rIPL produced minimal change in load cost (Low-High) or filter cost (Low-Filter), there was a change in the correlation pattern of the two measures across individuals. The measures were strongly correlated following control stimulation. Following TBS to either area, these correlations were reduced for contralateral targets, but remained unchanged for ipsilateral targets. Thus, while TBS of either region affected performance for contralateral targets, we did not observe dissociable contributions of MFG and IPL to spatial WM.
Subjects were tested on two sessions in a double blind behavioral study. In one session, subjects were given 1.25 mg of bromocriptine, a DA agonist, and in the other session, a placebo (PLAC). Using the design of McNab (2008), subjects performed a spatial WM task. On each trial, subjects viewed a display composed of 14 squares arranged along a virtual circle about a fixation point. Three different conditions were verbally cued at the start of each trial. In the "Low Load" condition, subjects were instructed to remember the location of 4 red circles. In the "High Load" condition, the memory load was increased to six items, 4 red and 2 yellow circles. In the "Filter" condition, 4 red and 2 yellow circles were presented but subjects were instructed to only remember the locations marked with red squares. The memory set was presented for 200 ms, followed by a 2 s delay period, followed by a probe, either within one of the to-be-remembered locations, or at a non-target location. Subjects indicated whether the probed location was at one of the target locations.
Subjects were divided into two groups, based on their performance on their spatial WM span. The High Span group was more accurate than the Low Span group after placebo (78% vs. 72%). Bromocriptine had minimal effects on accuracies, but larger RT effects were observed. Across all three WM conditions, the High Span group showed RT increases on bromocriptine as compared to placebo. In contrast, the Low Span group showed RT decreases on Bromocriptine. Both groups showed attenuated load and filtering costs after bromocriptine administration. Indeed neither group showed a RT increase on the Filter trials compared to the Low Load trials, suggesting that the dopamine either taxed performance on the Low Load trials or led to near-perfect filtering. Our results suggest two conclusions regarding DA and WM function. First, consistent with previous studies demonstrating an inverted U-shaped function relating dopamine and WM performance, high and low span individuals were affected in opposite directions by DA augmentation. Second, DA may reduce costs associated with increased demands on WM.
Participants made reaching movements to visual targets without vision of the limb. If the movement intersected the target, the target turned green and a pleasant “ding” was played. If the movement missed the target, the target turned red and an aversive “buzz” was played. A financial bonus system was used to increase motivation given the challenges of this categorical form of feedback. During a series of 441 reaches, a visuomotor rotation was gradually introduced between the “virtual” cursor location and the hand location, reaching a maximum of 8 degrees displacement (Izawa and Shadmehr, 2011). At the end of this adaptation phase, participants completed another 49 trials without any feedback. This washout phase allowed us to measure the temporal decay of motor learning. Participants were assigned to one of three groups. In the Success group, a TMS pulse was delivered on trials in which the reach intersected the target. In the Failure group, a TMS pulse was delivered on trials in which the reach missed the target. In the Control group, no TMS pulses were delivered.
The Failure and Control groups showed similar levels of learning during the adaptation phase, as well as similar decay during the washout phase. In contrast, the Success group did not learn the rotation and, as such, showed no aftereffect. While additional controls are required to rule out non-specific hypotheses (e.g., TMS is aversive, independent of the targeted location), the results suggest that disruption over motor cortex at the time of reward impairs motor adaptation. This would be consistent with the hypothesis that during normal learning, reward-signals following successful movements facilitate consolidation processes.
Participants learned to associate 16 abstract images with one of two response categories. Each trial had a 50% chance of being eligible for informative feedback. If ineligible, the response was followed by the feedback message “Unknown”. In the positive feedback condition, the feedback message “Correct” was displayed if the trial was eligible for feedback and the response was correct; otherwise the message “Unknown” was shown. Similarly, the message “Incorrect” was presented on eligible trials in the negative feedback condition when the answer was incorrect. In this manner, we separated positive and negative feedback-based learning. Participants were tested in two sessions in a double blind study. In one session participants ingested 1.25 mg of Bromocriptine (BR), a DA agonist; in the other session, they were given a placebo (PLAC).
On placebo, performance was more accurate in the positive condition (90%) than in the negative condition (83%). Overall, there was a Drug x Reinforcement Type interaction: BR resulted in a drop in accuracy for the positive condition and an increase in accuracy for negative condition. The reduction in performance from BR during positive reinforcement was similar for all alleles of COMT. In contrast, the benefit from BR during negative reinforcement was restricted to those individuals with higher levels of intrinsic DA (Val/Met or Met/MET alleles). For DRD2, the interaction of BR and reinforcement type was most pronounced in individuals with lower levels of intrinsic DA (A/A or A/G). These results suggest that individuals with higher DA levels in the frontal lobe may benefit in learning from negative feedback when DA signaling is increased, while individuals with lower striatal levels of DA may benefit in learning from both types of feedback with increased DA signaling.
EFFECTS OF THETA BURST STIMULATION OF MIDDLE FRONTAL GYRUS AND INFERIOR PARIETAL LOBULE ON LOAD AND FILTERING PROCESSES IN SPATIAL WORKING MEMORY
Peter A Butcher1, Richard B Ivry1;
1Department of Psychology, UC Berkeley
The exclusion of task-irrelevant information from working memory has been associated with activation in the middle frontal gyrus (MFG) and basal ganglia (McNab, 2007). In contrast, activation in right inferior parietal lobule (IPL) is associated with WM load. To further explore these hypotheses, we disrupted MFG and IPL in the right hemisphere with offline theta burst stimulation (TBS). On each trial, subjects viewed a display of 14 squares arranged along a virtual circle about a fixation point. Three different conditions were verbally cued for each trial. Low Load: Four red circles indicated possible target location. High Load: Memory load was increased to six items, 4 red and 2 yellow circles. Filter: 4 red and 2 yellow circles were presented but yellow circles were ignored. After a delay, a probe appeared and participants indicated if it was one of the target locations. While TBS to rMFG and rIPL produced minimal change in load cost (Low-High) or filter cost (Low-Filter), there was a change in the correlation pattern of the two measures across individuals. The measures were strongly correlated following control stimulation. Following TBS to either area, these correlations were reduced for contralateral targets, but remained unchanged for ipsilateral targets. Thus, while TBS of either region affected performance for contralateral targets, we did not observe dissociable contributions of MFG and IPL to spatial WM.
Subjects were tested on two sessions in a double blind behavioral study. In one session, subjects were given 1.25 mg of bromocriptine, a DA agonist, and in the other session, a placebo (PLAC). Using the design of McNab (2008), subjects performed a spatial WM task. On each trial, subjects viewed a display composed of 14 squares arranged along a virtual circle about a fixation point. Three different conditions were verbally cued at the start of each trial. In the "Low Load" condition, subjects were instructed to remember the location of 4 red circles. In the "High Load" condition, the memory load was increased to six items, 4 red and 2 yellow circles. In the "Filter" condition, 4 red and 2 yellow circles were presented but subjects were instructed to only remember the locations marked with red squares. The memory set was presented for 200 ms, followed by a 2 s delay period, followed by a probe, either within one of the to-be-remembered locations, or at a non-target location. Subjects indicated whether the probed location was at one of the target locations.
Subjects were divided into two groups, based on their performance on their spatial WM span. The High Span group was more accurate than the Low Span group after placebo (78% vs. 72%). Bromocriptine had minimal effects on accuracies, but larger RT effects were observed. Across all three WM conditions, the High Span group showed RT increases on bromocriptine as compared to placebo. In contrast, the Low Span group showed RT decreases on Bromocriptine. Both groups showed attenuated load and filtering costs after bromocriptine administration. Indeed neither group showed a RT increase on the Filter trials compared to the Low Load trials, suggesting that the dopamine either taxed performance on the Low Load trials or led to near-perfect filtering. Our results suggest two conclusions regarding DA and WM function. First, consistent with previous studies demonstrating an inverted U-shaped function relating dopamine and WM performance, high and low span individuals were affected in opposite directions by DA augmentation. Second, DA may reduce costs associated with increased demands on WM.