Portal de Periódicos da CAPES

Pondering the Pulvinar

2016; Cell Press; Volume: 89; Issue: 1 Linguagem: Inglês

10.1016/j.neuron.2015.12.022

1097-4199

Péter Lakatos, Monica N. O’Connell, Annamaria Barczak,

Neuroscience and Music Perception

While the function of the pulvinar remains one of the least explored among the thalamic nuclei despite occupying the most thalamic volume in primates, it has long been suspected to play a crucial role in attentive stimulus processing. In this issue of Neuron, Zhou et al., 2016Zhou H. Schafer R.J. Desimone R. Neuron. 2016; 89 (this issue): 209-220Abstract Full Text Full Text PDF PubMed Scopus (159) Google Scholar use simultaneous pulvinar-visual cortex recordings and pulvinar inactivation to provide evidence that the pulvinar is essential for intact stimulus processing, maintenance of neuronal oscillatory dynamics, and mediating the effects of attention. While the function of the pulvinar remains one of the least explored among the thalamic nuclei despite occupying the most thalamic volume in primates, it has long been suspected to play a crucial role in attentive stimulus processing. In this issue of Neuron, Zhou et al., 2016Zhou H. Schafer R.J. Desimone R. Neuron. 2016; 89 (this issue): 209-220Abstract Full Text Full Text PDF PubMed Scopus (159) Google Scholar use simultaneous pulvinar-visual cortex recordings and pulvinar inactivation to provide evidence that the pulvinar is essential for intact stimulus processing, maintenance of neuronal oscillatory dynamics, and mediating the effects of attention. It is customary to think of the thalamus as the switching station of the brain, through which cables carrying information from the outside world are switched over to cortex-bound cables where the information is then deciphered. In recent years, this simple view has undergone significant changes, driven in part by anatomical studies showing that thalamic relay nuclei consist of two types of thalamocortical projection systems: topographically organized, specific projections originating in the "core" regions of thalamic nuclei, and more diffusely projecting nonspecific projections originating in the thalamic "matrix" (Jones, 1998Jones E.G. Neuroscience. 1998; 85: 331-345Crossref PubMed Scopus (341) Google Scholar). While specific inputs carry detailed sensory information and project to the granular layer—the classical input layer—of primary sensory cortices, nonspecific modulatory projections predominantly target the supragranular layers of a wider array of cortical areas and are thought to be responsible for aligning internal excitability patterns to the timing of relevant sensory inputs, providing the neuronal context in which the hierarchically transmitted content, or specific information, is processed (Lakatos et al., 2013Lakatos P. Musacchia G. O'Connel M.N. Falchier A.Y. Javitt D.C. Schroeder C.E. Neuron. 2013; 77: 750-761Abstract Full Text Full Text PDF PubMed Scopus (286) Google Scholar). Another important anatomical finding instigating the changing view of the thalamus is that a number of thalamic nuclei receive a dominant part of their input not through pathways ascending from sensory organs, but from the cortex, for routing to other cortical or subcortical structures (Guillery and Sherman, 2002Guillery R.W. Sherman S.M. Neuron. 2002; 33: 163-175Abstract Full Text Full Text PDF PubMed Scopus (466) Google Scholar). The largest of these nuclei is the pulvinar, which occupies approximately a third of thalamic volume in humans and most non-human primates but is relatively miniscule in rodents (Chalfin et al., 2007Chalfin B.P. Cheung D.T. Muniz J.A. de Lima Silveira L.C. Finlay B.L. J. Comp. Neurol. 2007; 504: 265-274Crossref PubMed Scopus (47) Google Scholar). One could therefore argue that the pulvinar is likely to be involved in brain operations that are more elaborate in higher-order mammals. However, while the anatomy of the primate pulvinar is reasonably well described (Arcaro et al., 2015Arcaro M.J. Pinsk M.A. Kastner S. J Neurosci. 2015; 35: 9848-9871Crossref PubMed Scopus (95) Google Scholar), there is very little information about its function. Nonetheless, previous research indicates that the pulvinar is indeed involved in high-level functions like attention, social cognition, and speech processing. Thus, unsurprisingly, impaired pulvinar function is implicated in numerous psychiatric disorders including schizophrenia and ADHD (Benarroch, 2015Benarroch E.E. Neurology. 2015; 84: 738-747Crossref PubMed Scopus (70) Google Scholar). The clinical relevance and functionally significant role of the pulvinar in cognitive processes highlights the crucial role that non-human primate studies can play in enabling the mechanistic analysis of the multi-faceted pulvinar function. The study by Zhou et al., 2016Zhou H. Schafer R.J. Desimone R. Neuron. 2016; 89 (this issue): 209-220Abstract Full Text Full Text PDF PubMed Scopus (159) Google Scholar published in this issue of Neuron takes a significant step in the direction of determining the role of the pulvinar in attentive stimulus processing. First, in order to directly compare response properties and test for interactions in sites with overlapping receptive fields, the authors simultaneously recorded neuronal activity in ventro-lateral pulvinar and visual cortex (V4 and IT) in awake monkeys. A comparison of response properties showed that pulvinar receptive field size and stimulus selectivity is intermediate between V4 and IT cortex, consistent with its proposed role of transmitting visual information from lower- to higher-level nodes of the visual information processing hierarchy. Next, the authors examined the effects of spatial attention and found that both V4 and pulvinar responses were enhanced to attended versus unattended stimuli. Somewhat surprisingly, they discovered that attention effects in the pulvinar occurred later and were smaller in magnitude. While the former finding is in line with the notion of the pulvinar being positioned at a higher hierarchical stage than V4, the latter finding somewhat contradicts this notion as attention-related response amplitude changes tend to increase in downstream processing nodes (Cook and Maunsell, 2002Cook E.P. Maunsell J.H. J. Neurosci. 2002; 22: 1994-2004PubMed Google Scholar). Since one of the established effects of attention is an enhanced synchrony in the gamma frequency band, presumably for the facilitation of neuronal communication (Fries, 2015Fries P. Neuron. 2015; 88: 220-235Abstract Full Text Full Text PDF PubMed Scopus (1207) Google Scholar), the authors tested for synchrony within and across pulvinar and cortical neuronal ensembles by measuring the coherence of spiking activity and local field potentials (LFPs). While spike-LFP coherence was indeed enhanced within and across cortical sites in the gamma band when attention was directed to a location that fell within the receptive fields of V4 and IT neuronal ensembles, pulvinar spike-LFP coherence remained unchanged. Nevertheless, pulvinar-V4 spike-LFP gamma coherence increased with attention, and all gamma coherence effects were coupled with a decrease in coherence at lower frequencies. This indicates that attention affects neuronal synchrony between pulvinar and cortex similarly to what has been described earlier between cortical regions. In order to investigate the directionality of gamma frequency band interactions between pulvinar and cortex, the authors performed Granger causality analyses and found that while attention increased V4 influence on pulvinar, it had an opposite effect on the reverse interaction. These results are consistent with a V4 to pulvinar directed gamma band influence, which was substantiated by the additional finding that V4 gamma phase consistently led pulvinar gamma. Notably, pulvinar had a greater causal influence on V4 neuronal ensemble activity at lower frequencies, and consistent with a previous study (Saalmann et al., 2012Saalmann Y.B. Pinsk M.A. Wang L. Li X. Kastner S. Science. 2012; 337: 753-756Crossref PubMed Scopus (595) Google Scholar), when anticipating the cued stimulus, this influence was enhanced by attention at alpha frequencies. These opposite direction effects at gamma versus alpha frequencies likely reflect bottom-up, stimulus-bound versus top-down, internally driven (e.g., anticipation) attentional influences, which would fit with data suggesting that alpha frequency band oscillations are involved in orchestrating neuronal synchrony for top-down, whereas gamma oscillations are utilized for bottom-up, neuronal communication (Fries, 2015Fries P. Neuron. 2015; 88: 220-235Abstract Full Text Full Text PDF PubMed Scopus (1207) Google Scholar). To further investigate pulvinar-cortex interactions, Zhou et al., 2016Zhou H. Schafer R.J. Desimone R. Neuron. 2016; 89 (this issue): 209-220Abstract Full Text Full Text PDF PubMed Scopus (159) Google Scholar employed pulvinar deactivation with simultaneous visual cortex recordings while the monkeys performed the spatial attention task. As expected, the subjects' performance degraded significantly with pulvinar deactivation. Importantly, performance was affected only in the portion of the visual field that corresponded to the inactivated region of the pulvinar. The attention-related enhancement of V4 responses was also greatly reduced, to a level even below responses to ignored stimuli, indicating that the cortical response degradation was at least partly due to a loss of responsiveness of V4 neuronal ensembles or a reduction in the input to V4. Since the authors did not detect a change in the response of LGN neurons during deactivation, they concluded that the pulvinar is not only needed in order to enhance cortical stimulus processing when stimuli are attended, but it is also part of the circuitry responsible for intact information processing. We speculate that one explanation for the authors' finding is that V4 and other cortical areas reciprocally connected to the pulvinar act as coincidence detectors: if the information transmitted to the pulvinar is not projected back strongly, responses in the cortex are diminished. If true, this could serve as a very powerful suppressive mechanism in at least 2 ways: either the pulvinar itself could be suppressed by the thalamic reticular nucleus (TRN) so that it cannot project back and "confirm" cortical responses (similarly to pulvinar inactivation), or the spatiotemporal structure of information (neuronal responses) received from the cortex could be modified and thus, when back-projected to cortex, would not result in coincidence detection (confirmation). Accompanying the reductions in responses to attended stimuli, pulvinar deactivation also resulted in a robust reduction of spike-LFP coherence in the gamma band within V4, and gamma coherence between pulvinar and V4. Interestingly, these effects were accompanied by a large increase in the amplitude of lower-frequency oscillations. Taken together, the authors argue that the pulvinar is needed to maintain visual cortex in an active state, which enables both responding to sensory inputs and top-down modulation of cortical responses by other structures like the frontal eye fields (Buschman and Miller, 2007Buschman T.J. Miller E.K. Science. 2007; 315: 1860-1862Crossref PubMed Scopus (1599) Google Scholar). The pulvinar's role as an enabler, rather than a direct mediator of cortical function, is supported by both the timing of attention effects and the interactions at gamma frequencies, which show a consistent V4 lead. It is likely that similar to cortical operations, pulvinar's machinery subserving attentional operations is supervised by the prefrontal cortex. However, certain regions of the pulvinar, like the ventrolateral region, do not receive inputs from prefrontal cortex (Zhou et al., 2016Zhou H. Schafer R.J. Desimone R. Neuron. 2016; 89 (this issue): 209-220Abstract Full Text Full Text PDF PubMed Scopus (159) Google Scholar); thus, the supervisor needs a mediator structure that, by adjusting pulvinar's neuronal dynamics, can convey the message of current attentional goals to sensory cortices. A likely candidate for this mediator role is the TRN, which is extensively connected to prefrontal regions and the pulvinar (Zikopoulos and Barbas, 2006Zikopoulos B. Barbas H. J. Neurosci. 2006; 26: 7348-7361Crossref PubMed Scopus (259) Google Scholar). Intriguingly, the TRN itself receives inputs from most sensory cortical regions but, unlike the pulvinar, does not have reciprocal connections. However, it extensively innervates thalamic relay nuclei. Similar to the prospective role of the pulvinar, the prefrontal cortex-TRN-thalamic relay nuclei circuitry was also proposed to play a role in selective attention by modulating the strength of non-specific thalamocortical inputs, originating in the thalamic matrix, that go on to modulate excitability in sensory cortical regions (Lakatos et al., 2013Lakatos P. Musacchia G. O'Connel M.N. Falchier A.Y. Javitt D.C. Schroeder C.E. Neuron. 2013; 77: 750-761Abstract Full Text Full Text PDF PubMed Scopus (286) Google Scholar, Zikopoulos and Barbas, 2006Zikopoulos B. Barbas H. J. Neurosci. 2006; 26: 7348-7361Crossref PubMed Scopus (259) Google Scholar), which recently received some experimental evidence (Wimmer et al., 2015Wimmer R.D. Schmitt L.I. Davidson T.J. Nakajima M. Deisseroth K. Halassa M.M. Nature. 2015; 526: 705-709Crossref PubMed Scopus (279) Google Scholar). So what might the purpose of the two different routes—prefrontal-TRN-relay nuclei and prefrontal-TRN-pulvinar—be in attentive stimulus processing? We hypothesize that the former network prioritizes active sensing by aligning processing resources to an external temporal reference frame derived from the temporal structure of inputs related to attended external events. Conversely, the latter route provides an internal temporal reference frame utilizing alpha band oscillations in order to facilitate communication between neuronal ensembles processing attended stimulus features (Saalmann et al., 2012Saalmann Y.B. Pinsk M.A. Wang L. Li X. Kastner S. Science. 2012; 337: 753-756Crossref PubMed Scopus (595) Google Scholar) and, as Zhou et al., 2016Zhou H. Schafer R.J. Desimone R. Neuron. 2016; 89 (this issue): 209-220Abstract Full Text Full Text PDF PubMed Scopus (159) Google Scholar suggest, maintain task-relevant cortical regions in an active state. To summarize, the study by Zhou et al., 2016Zhou H. Schafer R.J. Desimone R. Neuron. 2016; 89 (this issue): 209-220Abstract Full Text Full Text PDF PubMed Scopus (159) Google Scholar provides causal information for the involvement of the pulvinar in attentive visual stimulus processing and in maintaining and modulating neuronal oscillatory dynamics in visual cortex. Their results also clearly demonstrate that the influences the pulvinar exerts on cortical operations have simultaneous bottom-up and top-down signatures. We propose that a possible explanation for this is that the authors' data reflect the workings of two distinct core-matrix organized thalamocortically projecting systems in the brain: one bottom-up-like, focused on the external, and one top-down-like, focused on the internal world (Figure 1). In this theoretical framework, while core regions in both systems would transmit specific external versus internal information, accompanying matrix ensembles would be responsible for the timing of cortical activity in task-relevant regions in support of effective communication and, subsequently, the enhancement and merging of relevant information. The external core-matrix system consists of thalamocortical sensory relay nuclei (Jones, 1998Jones E.G. Neuroscience. 1998; 85: 331-345Crossref PubMed Scopus (341) Google Scholar) and utilizes highly malleable delta-theta frequency band neuronal oscillations aligned to the temporal structure of relevant sensory inputs for the timing of cortical operations (Lakatos et al., 2013Lakatos P. Musacchia G. O'Connel M.N. Falchier A.Y. Javitt D.C. Schroeder C.E. Neuron. 2013; 77: 750-761Abstract Full Text Full Text PDF PubMed Scopus (286) Google Scholar). The internal core-matrix system, which includes the pulvinar, would utilize self-timed and comparatively more rigid alpha frequency band oscillations to synchronize neuronal activity in relevant cortical regions. These two systems would both be orchestrated by the TRN with the supervision of prefrontal cortex. While further experiments are needed to evaluate this hypothesis, the results of Zhou et al., 2016Zhou H. Schafer R.J. Desimone R. Neuron. 2016; 89 (this issue): 209-220Abstract Full Text Full Text PDF PubMed Scopus (159) Google Scholar, together with previous studies (Purushothaman et al., 2012Purushothaman G. Marion R. Li K. Casagrande V.A. Nat. Neurosci. 2012; 15: 905-912Crossref PubMed Scopus (149) Google Scholar), leave no doubt that cortical processing can only be understood if the role of the pulvinar is pondered. Pulvinar-Cortex Interactions in Vision and AttentionZhou et al.NeuronJanuary 06, 2016In BriefThe pulvinar is often proposed to modulate cortical processing with attention. Zhou et al. find that beyond any role in attention, the pulvinar input to cortex seems necessary to maintain the cortex in an active state. Full-Text PDF Open Archive