Trial-by-Trial Dynamics: A Window in Time

The arrow of time is of little consequence to the majority of physics. Except in a few cases, the laws of physics still apply whether one observes them in a reverse or in a forward manner. The contrasting notion relevant to this discussion is that our experience of time flows uni-directionally into the future. If our experimental approaches manifest as a function of our experience, then it has been an unsurprising technique in Cognitive Neuroscience to observe the consequences of an act of cognition in the direction of time’s arrow: i.e., to observe the neural correlates of cognition after an event of interest. With few exceptions (Remington, 1969; Squires et al., 1976) this influential approach has characterized most of cognitive neuroscience. The recent paper by Eichele et al. (2010) represents a methodological reversal of time’s arrow to illustrate the conditions that precede a cognitive event of interest. The authors’ examine how brain activity under the control of a previously established performance monitoring network can serve to define an error-prone brain state well in advance of an imminent error. This approach offers insight into the dynamic and adaptive properties of the brain that would otherwise be relegated to the status of unexplained variance and gives an avenue for possible insight into causal, rather than associated mechanisms of perception and action. A classic approach within human electrophysiology has been to link the neural consequences of cognition to an averaged event-related signal that differs as a function of an a priori experimental manipulation. The most conventional approach here has been to use averaged eventrelated responses to characterize the brain’s transient response to the event. However, the neuronal activity prior to the onset of the event exhibits large fluctuations on time scales of hundreds of milliseconds to seconds. As such, a burgeoning direction of research has examined how this pre-stimulus activity influences perception, cognition, and behavior. The underlying hypothesis here is that the fluctuations in pre-stimulus neural activity reflect brain states that influence the moment-to-moment processing, as well as later cognitive and motor decisions and outputs. The majority of the studies in this fledgling field (Eichele et al., 2010, being a notable exception) have largely focused on the spontaneous variability of neuronal fluctuations indexed by oscillatory activity. Several recent studies point to visual and somato-sensory perception being modulated by pre-stimulus oscillatory activity. In particular, activity in the alpha band (∼10 Hz) has been shown to predict failures in visual (Ergenoglu et al., 2004; Thut et al., 2006; van Dijk et al., 2008) and somato-sensory perception (Linkenkaer-Hansen et al., 2004). Moreover, fluctuations in alpha activity have also been found to predict failures in sustained attention (Mazaheri et al., 2009; O’Connell et al., 2009). The work of Eichele represents an exceptional extension of this approach into a clever examination of the pre-trial behavioral and stimulus contingencies as related to the evolution of the evoked response and the subsequent commission of an error. This approach manifests primarily through the re-operationalization of an experimental condition as a function of stimulus contingencies and subsequent behavior rather than an a priori-experimental manipulation. If our goal is to model neural activity at the systems level as a function of cognition and behavior, a combination of the classic approach with this trial by trial dynamic approach would prove a powerful avenue by which to pursue the stubborn question of how the brain, perception, and behavior interact. An interesting question for future research arises as a consequence of Eicheles’ and others’ work concerning trial by trial dynamics of perception and behavior relates directly to the work of Squires referenced in the quote at the beginning of Eichele et al.: