Medial frontal cortical neurons code errors made by others

Learning from the errors of others’ is one of the most fundamental of cognitive abilities, as so well captured by the phrase “The wise learn by the mistakes of others, fools by their own”. There is neuroimaging evidence pointing to medial frontal cortical areas as a candidate region that makes it possible to learn from the mistakes of others, however, there have been a number of open questions, including the precise loci of observed-error processing, whether or not self-generated errors and those made by others are processed by the same neurons, and how the neural mechanisms of observed-error processing shape one’s own behavior.

In their recent study, Yoshida et al. (2012) investigated, using neurophysiological single-cell recordings, how medial frontal cortical neurons fire when macaque monkeys observe errors of another monkey. Two monkeys facing one another alternated in a choice task and, indicating that the monkeys were learning from each other’s mistakes, the monkeys correctly guided their own choice in most trials subsequent to no-reward trials of the other monkey. Cells firing during observation of another’s error were found in two medial frontal regions (convexity and sulcus), and about half of these neurons fired only when observing other’s errors (and not also during errors committed by oneself). The authors further suggested that the convexity subregion is more specifically involved in detection of others’ errors, and that the sulcus subregion is more important for guiding one’s behavior based on the errors committed by others.

These highly interesting findings shed light on the neural mechanisms underlying observational learning, demonstrate that there are neurons specifically responding to mistakes made by others and that there is fine regional specialization supporting error detection and shaping of one’s subsequent behavioral choices. This study also presents a very nice example of how specific aspects of a behavioral task can be isolated and associated with specific aspects of neural activity

Reference: Yoshida K, Saito N, Iriki A, Isoda M. Social error monitoring in macaque frontal cortex. Nature Neuroscience (2012) advance online publication http://dx.doi.org/doi:10.1038/nn.3180


Superior temporal sulcus as the hub for distributed brain networks that support social perception

Over the last decade, there has been a significant surge of interest towards the neural underpinnings of social cognition, and indeed several candidate brain structures have been implicated based on the results of such studies. While the vast majority of these studies have utilized impoverished stimuli and task paradigms (e.g., contrasting two perceptual categories such as faces vs. bodies), there have been recently also studies that have utilized naturalistic stimuli such as movie clips to investigate the neural basis of social cognition.

In their recent study, Lahnakoski et al. (2012) presented, during 3-Telsa functional magnetic resonance imaging, healthy volunteers with a collection of short movie clips containing both social features (i.e., faces, human bodies, biological motion, goal-oriented actions, emotions, social interactions, pain, and speech) and non-social features (i.e., places, objects, rigid motion, people not engaged in social interaction, non-goal-oriented action, and non-human sounds). Brain activity patterns were then modeled based on the time course of occurrence of these social and non-social features.

Interestingly, the authors observed that the posterior superior temporal sulcus responded to all social features but not to any of the non-social features. Furthermore, there were four extended networks that participated in processing of specific social signals: 1) a fronto-temporal network responding to multiple social categories, 2) a fronto-parietal network preferentially activated by bodies, motion, and pain, 3) a temporal-lobe-amygdala network responding to faces, social interaction, and speech, and, finally, 4) a fronto-insular network that activated during perception of emotions, social interactions, and speech. Taken together, these results disclose the posterior superior temporal sulcus as a central hub for distributed brain networks that support social perception, and add to accumulating pool of evidence indicating that utilization of naturalistic stimuli in fMRI studies provides an effective tool for the study of the neural basis of social cognition.

Reference: Lahnakoski JM, Glerean E, Salmi J, Jääskeläinen IP, Sams M, Hari R, Nummenmaa L. Naturalistic fMRI mapping reveals superior temporal sulcus as the hub for distributed brain network for social perception. Frontiers in Human Neuroscience (2012) 6: 233. http://dx.doi.org/10.3389/fnhum.2012.00233


Cortical-cerebellar loops during attention to visual motion

For a long time the role of the human cerebellum was thought to be limited to motor functions such as coordination of well-learned motor sequences (e.g., an experienced golf player swinging the golf club). It is, however, being increasingly recognized in the neuroscience community that specific parts of the cerebellum play a much wider role in human cognitive functions than what has been previously assumed.

In a recent study by Kellermann et al. (2012), the involvement of cerebellar-cortical loops in a visual attention-to-motion task was investigated using functional magnetic resonance imaging. Healthy volunteers were shown stationary vs. moving grating stimuli; in the test condition the subjects were instructed to attend slight changes in the speed that the bars were moving. In reality there were no changes in movement velocity and thus the only factor that was experimentally manipulated in the test vs. control conditions was the level of attention to visual motion.

The authors observed increased effective connectivity between the cerebellum and neocortical dorsal visual stream structures with increasing level of attention to visual motion. Further, it was observed that, under attention, functional connectivity from cerebellum to visual area V5 (that processes visual motion) was enhanced, whereas connectivity from V5 to posterior parietal cortex (that is a higher-order attention-directing structure in the brain) was attenuated.

The authors interpret these findings as indicating that, under conditions where visual motion is highly predictable (i.e., when internal models can strongly guide perception), the posterior parietal cortex feeds top-down predictions to the hierarchically lower motion processing area V5 via crus I of cerebellum (thus potentially enhancing the precision of input-predictions of V5 neurons), while at the same time influence of bottom-up inputs from V5 to posterior parietal cortex are suppressed; the authors further note that the task-specific input-output patterns of the cerebellum likely determine the functional role that the cerebellum plays in various cognitive processes. Overall, these findings highlight the importance of cerebellar-cortical loops in perceptual-cognitive functions, something that has been regrettably often neglected in the human functional neuroimaging literature.

Reference: Kellermann T, Regenbogen C, De Vos M, Mößnang C, Finkelmeyer A, Habel U. Effective connectivity of the human cerebellum during visual attention. Journal of Neuroscience (2012) 32: 11453–11460. http://dx.doi.org/ 10.1523/JNEUROSCI.0678-12.2012


Presenting a rehearsed melody during slow-wave sleep enhances learning of the melody

It is increasingly recognized that learning of skills is facilitated by sleep. Intermittent sleep, even for briefer periods (napping), leads to increasing level of performance on a task that has been rehearsed prior to sleeping. Furthermore, the role of memory consolidation during sleep has been observed to depend on whether one sees task-related dreams; in studies where subjects have been awakened in the middle of sleep and requested to recall their dream contents, task-related dream content predicted higher post-sleep increments in task performance.

In their recent study, Antony et al. (2012) studied whether one could facilitate learning of skills by external stimulation related to task learning that does not wake up the skill-learner. Specifically, the authors hypothesized that the ability to produce a melody could be influenced by auditory cuing during sleep. Volunteers practiced two melodies for an equal amount of time. During an afternoon nap following the training session, one of the melodies was presented during slow-wave sleep detected with electroencephalography. Post-sleep testing revealed that performance of the melody that was played during slow-wave sleep was better than performance of the other melody (prior to the nap there were no differences in performance of the two melodies). Performance enhancements further correlated with the amount of time that the subjects were under slow-wave sleep.

These highly interesting results further contribute to a rapidly growing and exciting area of research in cognitive neuroscience on the importance of sleep for memory consolidation and learning of skills. These results further underline the importance of sleep for learning and suggest that it is possible to facilitate the beneficial effects of sleep on learning of musical sequences by external stimulation during a specific period of sleep, the slow wave sleep.

Reference: Antony JW, Gobel EW, O’Hare JK, Reber PJ, Paller KA. Cued memory reactivation during sleep influences skill learning. Nature Neuroscience (2012) 15: 1114-1116. http://dx.doi.org/10.1038/nn.3152


Functional connectivity of dorsal vs. ventral posterior parietal cortices during top-down vs. bottom-up attention to memory

In previous neuroimaging studies, the human posterior parietal cortex has been identified as one intimately involved in attentional processes. Dorsal aspects of the posterior parietal cortex have been associated with “top-down” attention (i.e., focusing attention on external stimuli based on internal goals of the subject) and ventral aspects of the posterior parietal cortex have been associated with “bottom-up” attention (i.e., externally presented unexpected stimuli capturing one’s attention). Whilst studying attention to externally applied stimuli is experimentally convenient, focusing attention on internal events (e.g., conducting memory searches) is equally important. One might assume that the posterior parietal cortex governs attention to memorized items similarly as in the case of externally applied stimuli, however, it has not been systematically tested whether the posterior parietal cortex is functionally segregated into dorsal and ventral areas during cued vs. non-cued recognition memory trials.

In their recent study, Burianová et al. (2012) tested, by presenting cued vs. non-cued recognition memory trials during functional magnetic resonance imaging, whether dorsal aspects of the posterior parietal cortex are more involved in top-down memory searches and whether ventral aspects of the posterior parietal cortex are more involved in non-cued bottom-up recognition memory. The authors observed spatially dissociable networks of brain areas that overlapped only in precuneus. During cued recognition memory trials (“top-down”), dorsal posterior parietal cortex was functionally connected with areas comprising the dorsal attention network as well as with memory-related brain areas; there was further a significant correlation between cued memory recognition performance and this network activity. In contrast, during uncued trials, ventral posterior parietal cortex was functionally connected with the ventral attention system and with relevant memory areas. These findings thus disclose a nice double-dissociation of roles between dorsal and ventral posterior parietal cortical areas in recognition memory that closely resembles the distinct roles that these areas play in top-down vs. bottom-up attention.

Reference: Burianová H, Ciaramelli E, Grady CL, Moscovitch M.  Top-down and bottom-up attention-to-memory: mapping functional connectivity in two distinct networks that underlie cued and uncued recognition memory. Neuroimage (2012) e-publication available prior to publication. http://dx.doi.org/10.1016/j.neuroimage.2012.07.057