The Miller Lab

  • Robust Gamma Coherence between Macaque V1 and V2 by Dynamic Frequency Matching

    Miller Lab

    Rhythmic synchrony between neurons has been suggested as a mechanism for establishing communication channels between neurons.  However, this hypothesis has been criticized because of observations that the exact frequency of gamma oscillations bounces around too much to provide a stable communication channel.  (BTW, it doesn’t seem to bother anyone that single neuron activity also bounces around).

    In this study, Roberts et al record from V1 and V2 simultaneously while presenting gratings of varying contrast.  Even though the gamma frequencies changed with stimulus contrast and fluctuated over time, coherence remained stable between V1 and V2.   Thus, rhythmic synchrony can provide a stable channel for neural communication.
    http://www.cell.com/neuron/abstract/S0896-6273(13)00227-4?utm_source=feedly

    For further reading on the role of rhythmic synchrony in neural communication see:
    Miller, E.K. and Buschman, T.J. (2013) Cortical circuits for the control of attention.  Current Opinion in Neurobiology.  23:216–222  View PDF »

    Buschman, T.J., Denovellis, E.L., Diogo, C., Bullock, D. and Miller, E.K. (2012) Synchronous oscillatory neural ensembles for rules in the prefrontal cortex. Neuron, 76: 838-846.  View PDF

  • A Functional Hierarchy within the Parietofrontal Network in Stimulus Selection and Attention Control

    Miller Lab

    Ibos et al examined the relative roles of the frontal eye fields (FEF) and lateral intraparietal area (LIP) in bottom-up vs top-down selection.  They found that intrinsic salience (bottom-up) was signaled in LIP before the FEF whereas extrinsic salience (top-down) was signaled in FEF before LIP.  The authors conclude that bottom-up vs top-down control of attention predominates in the parietal vs frontal cortex, respectively.
    http://www.jneurosci.org/content/33/19/8359.abstract

    As noted by the authors, this is highly consistent with our lab’s observations that attention signals for bottom-up capture by stimulus salience (pop-out) vs top-down search originate from parietal vs frontal cortex, respectively.
    Buschman, T.J. and Miller, E.K. (2007) Top-down versus bottom-up control of attention in the prefrontal and posterior parietal cortices. Science. 315: 1860-1862   View PDF »

    We also showed that rhythmicity of frontal cortical top-down signals may control the periodic shifts of attention during visual search that leads to eventual selection of a target:
    Buschman, T.J. and Miller, E.K. (2009) Serial, covert, shifts of attention during visual search are reflected by the frontal eye fields and correlated with population oscillations. Neuron, 63: 386-396. View PDF »

  • Neural Limits to Representing Objects Still within View

    Miller Lab

    Our very limited ability to hold multiple thoughts in mind is apparent to anyone who has tried to talk on the phone and email at the same time.  Traditionally, this is thought of as a limitation in the capacity of working memory, the “mental scratchpad” used to keep important information “online” after it is no longer available.  However, Ed Vogel and crew show that the bottleneck is not in working memory per se but instead present during processing of visual stimuli while they are still visible.  Thus, the bottleneck is not (just) in memory but also in the processing of sensory inputs.  In other words, capacity limitations seem to be a fundamental limit in neural processing related to consciousness in general, not a unique byproduct of working memory.
    http://www.jneurosci.org/content/33/19/8257.abstract

    As noted by the authors, this is consistent with our finding that when capacity is exceeded, information is lost in a bottom-up fashion during initial processing of visual stimuli:
    Buschman,T.J., Siegel, M., Roy, J.E. and Miller, E.K. (2011) Neural substrates of cognitive capacity limitations. Proceedings of the National Academy of Sciences. 108(27):11252-5. View PDF »

  • Multiplexing Stimulus Information through Rate and Temporal Codes in Primate Somatosensory Cortex

    Miller Lab

    Two different features of a vibrotactile stimulus are encoded by rate and temporal codes in primary sensory cortex.  The amplitude was reflected in the overall firing rate of neurons whereas the frequency was reflected in the oscillatory phase-locking of spikes.
    Harvey et al, 2013 PLOS Biology

    The two coding schemes, rate vs temporal codes, are often debated as if they are in opposition and mutually exclusive.  They are not and this paper elegantly demonstrates this important point.  And this is not just limited to vibratory tactile stimuli.  Schroeder and colleagues have argued that all sensory processing involves periodic sampling and rhythmic entrainment of cortical neurons.  (We, for example, periodically sample vision via periodic eye movements and shifts of attention.)

    Plus, rhythmic synchrony allows multiplexing, not only by adding another coding dimension as shown here, but also by allowing neurons to communicate different messages to different targets depending on whom they are synchronized with (and how, e.g., phase, frequency).  That way, the same neurons can participate in different functions yet still convey unambiguous messages.  For a brief discussion of this latter point and why we need multiplexing in the cortex see: Miller, E.K. and Fusi, S. (2013) Limber neurons for a nimble mind. Neuron. 78:211-213. View PDF

  • Cerebellum and Cognition: Evidence for the Encoding of Higher Order Rules

    Miller Lab
    , ,

    This paper reports FMRI in humans performing a task requiring first-order rules (S-R associations with a specific motor output) and second order rules that govern the use of the first-order rules.  Cerebellum lobules that project to the prefrontal cortex show activation for both types of rules.  This suggests that the cerebellum contributes to rule-based behaviors even when the rules are higher-order and don’t directly involve a motor command.
    http://cercor.oxfordjournals.org/content/23/6/1433.abstract

    For further reading on the role of rules in cognition and their neural implementation see:
    Miller, E.K. and Cohen, J.D. (2001) An integrative theory of prefrontal cortex function. Annual Review of Neuroscience, 24:167-202.  View PDF »

    Buschman, T.J., Denovellis, E.L., Diogo, C., Bullock, D. and Miller, E.K. (2012) Synchronous oscillatory neural ensembles for rules in the prefrontal cortex. Neuron, 76: 838-846.  View PDF

    Wallis, J.D., Anderson, K.C., and Miller, E.K. (2001) Single neurons in the prefrontal cortex encode abstract rules. Nature, 411:953-956. View PDF »

  • WSJ: How the Brain Really Works. (It’s not a phrenology map)

    Miller Lab
    , ,

    A nice article in the Wall Street Journal describing Jack Gallant’s recent FMRI work.  They didn’t just  subtract conditions and come up with a typical imaging map with one or a few isolated bits of activation.  Jack L. Gallant, Tolga Çukur and colleagues used sophisticated  analyses to find the relationship between the patterns of whole brain activity and the content of videos watched by the subjects.  This revealed wide networks, not isolated patches, of neurons engaged by attention to different things in the video (humans vs vehicles, etc).

    It also showed how dynamic and flexible the brain is.  When subjects looked for humans, large portions of the cortex were sensitive to humans and less sensitive to vehicles. When subjects looked for vehicles, large portions of the cortex became vehicle detectors.  Many of the same brain areas were involved in multiple networks, changing when people changed the focus of their attention.  Thus, rather than the cortex being composed of modules with strict specializations, high-level information is spread across wide-ranging cortical networks of neurons that participate in many different functions, adapting their properties to current cognitive demands.

    We have long argued that mixed selectivity, adaptive coding neurons are crucial for hallmarks of cognition like flexibility.  And in forthcoming paper (Rigotti et al), we show computationally that you can’t build a complex brain w/o them.

    For a brief discussion of this issue, read this Preview of a paper by Stokes et al:
    Miller, E.K. and Fusi, S. (2013) Limber neurons for a nimble mind. Neuron. 78:211-213. View PDF

    And stay tuned for this paper:
    Rigotti, M., Barak, O., Warden, M.R., Wang, X., Daw, N.D., Miller, E.K., & Fusi, S. (in press) The importance of mixed selectivity in complex cognitive tasks. Nature.