«A Thesis Presented in Partial Fulfillment of the Requirements for the Degree Master of Science Approved April 2014 by the Graduate Supervisory ...»
The Effects of Implied Motion Training on General Cortical Processing
A Thesis Presented in Partial Fulfillment
of the Requirements for the Degree
Master of Science
Approved April 2014 by the
Graduate Supervisory Committee:
Jose Nanez, Chair
Arizona State University
Current research has identified a specific type of visual experience that leads to faster cortical processing. Specifically, performance on perceptual learning of a directional-motion leads to faster cortical processing. This is important on two levels;
first, cortical processing is positively correlated with cognitive functions and inversely related to age, frontal lobe lesions, and some cognitive disorders. Second, temporal processing has been shown to be relatively stable over time. In order to expand on this line of research, we examined the effects of a different, but relevant visual experience (i.e., implied motion) on cortical processing. Previous fMRI studies have indicated that static images that imply motion activate area V5 or middle temporal/medial superior temporal complex (MT/MST+) of the visual cortex, the same brain region that is activated in response to real motion. Therefore, we hypothesized that visual experience of implied motion may parallel the positive relationship between real directional-motion and cortical processing. Seven subjects participated in a visual task of implied motion for 4 days, and a pre- and post-test of cortical processing. The results indicated that performance on implied motion is systematically different from performance on a dot motion task. Despite individual differences in performance, overall cortical processing increased from day 1 to day 4.
This Thesis would never have been written if not for the help of the following:
Dr. Jose E. Nanez Sr.
Daniel Zimmerman ii
TABLE OF CONTENTSPage Abstract
Acknowledgements….……………………………………………………………….…......ii Introduction………….………………………………………………………………..….....1 Methods
Material & Procedures
Footnotes….………………………………………………………………………………. 18 References…………………………………………………………………………….........19 Tables
Tables 1 - 3
Tables 4 - 7
Tables 8 & 9
Tables 10 & 11
Tables 12 - 14
Figures 1 - 3
Figures 4 & 5
Figures 6 & 7
Figures 8 & 9
Perceptual learning is broadly defined as improvement in sensory perception by way of training over time (Fahle, 2005). In addition, perceptual learning has been found to be highly specific to low-level features (i.e., contrast, direction, orientation, location, etc). For example, learning to discriminate motion in one location will not transfer to a location equal to and greater than 3 degrees away, which is within the primary visual cortex1. That is, neurons firing rate drops to baseline when location, direction, orientation of the stimuli is shifted (Fahle, 2005). Moreover, the high specificity suggests that cells in V1 integrate multiple low-levels features for more complex visual processing. For example, moving dots at 10 percent coherency (local motion) within randomly distributed dots will appear to move in one direction (global motion). Taken together, the integration suggests that V1 extracts great amount of information from single low-level features. The high specificity of perceptual learning has been empirically linked to a biological process of neuroplasticity, which is a functional (performance enhancement) and structural (increase in synaptic pathways) enhancement to novel changes in the environment. This neuro-biological process explains how we learn and adapt quickly and proficiently to new information in our environment (Fahle & Poggio, 2002).
Researchers have been exploring the generalizability of perceptual learning within low-level (e.g., change in location) and high-level visual processing (e.g., complex visual processing). For example, Crist, Kapadia, Westheimer, & Gilbert, (1997) found that performance in three-line bisection task2 transferred up to eight degrees from the training field, which is beyond the primary visual cortex. Furthermore, perceptual learning of coherent dot motion, where subjects discriminate direction of a specified group of dots
(high-level visual process). Very recent studies have indicated that perceptual learning of motion does not only lead to faster flicker perception, but also faster reading speed and better reading comprehension (Holloway, Nanez, & Seitz, under revision; Groth, (2013, unpublished MS thesis).
The dual system theory has been proposed to explain the systems of low- and high- level visual processing, and the interaction between the two levels (Milner & Goodale, 2007). According to the Milner and Goodale (2007), humans have two distinct, but mutually independent visual systems, which consist of visual streams for action (i.e., the dorsal stream) and perception (i.e., ventral stream). More specifically, when exposed to visual stimuli (i.e., moving dots), the information is carried from the ganglion cells within the retina to the magnocellualr pathways (large layer of cells within the lateral geniculate nucleus), then to the primary visual cortex for processing (V1). From V1, depending on the visual stimuli, the information travels through either the dorsal stream (“where pathway”) or the ventral stream (“what pathway”). If the stimulus is fast moving low-contrast dots, the information travels from V1 to the MT and parietal lobe for higherorder processing (e.g., where tasks like flicker perception, word decoding, reading comprehension are processed) via the dorsal stream. If subjects are instructed to identify an object that is presented at a slower rate with higher contrast, then the information is projected to the inferior temporal cortex through the ventral stream (Nealey & Maunsell, 1994).
In regards to our motion discrimination task, specifically, the neural properties of area MT provide the theoretical framework for the current study. Moreover, area MT has
decrease in neural activity when an observer is repeatedly exposed to the same direction of motion. In addition, area MT has demonstrated an increase in functional benefit when exposed to novel motion (Kohn, 2007; Ranganath & Rainer, 2003). Essentially, motion discrimination training that is randomly displayed throughout the experiment is structurally and functionally beneficial for area MT. In this paper, we are only interested in the neural correlates of V5 or medial temporal through the dorsal stream. The neural correlates of magnocallular pathway, MT/MST+, and the dorsal stream may provide the theoretical framework for the variables of interest in this paper. That is to say, the variables of interest include fast moving, low contrast, unidirectional-implied motion.
Each component of the variable is related to low-level and high-level visual processing.
Implied motion is broadly defined as static stimuli that imply motion due to dynamic features within the stimuli (Kourtzi & Kanwisher, 2000). For example, the image of a silhouette in Figure 1 is implying motion to the left with high leg-left, arched back, and extended arms. In contrast, the image of a woman in Figure 3 is motionless because she lacks the articulation in body posture needed to imply motion. Previous research has indicated that images resembling Figure 1 activate area MT/MST+ significantly higher than the images resembling Figure 3. Kourtzi and Kanwisher, (2000) argued that MT/MST+ is not the only brain region responsible for processing visual cues that imply motion; rather it is part of a larger network of brain regions that “mediate” the visual processing of motion. In addition, they reasoned that MT/MST+ sensitivity of implied motion are, in part, governed by a top-down influence. In other words, the adult brain has been exposed to similar animate objects (i.e., biological figure of a silhouette),
Resonance Imaging (fMRI) studies of implied motion following Kourtzi & Kanwisher, (2000) have confirmed the top-down influence, by showing neural correlates beyond MT/MST+. For example, researchers have found activation in superior temporal sulcus (STS; sensitive to dynamic biological forms), extrastriate body area (EBA; sensitive to body type), and pre-motor and motor cortex (sensitive to high degree of articulation) (Proverbio, Riva, & Zani, 2009; Jellema & Perrett, 2003). In addition, Fawcett, Hillebrand, & Singh, (2007) and Lorteije et al., (2006) found late activation (600 msms) in MT, suggesting the influence of higher-level processing. However, researchers also argue that activation of implied motion could be due to differences in low-level features (Lorteije et al., 2011; Pavan, Cuturi, Maniglia, Casco, & Campana, 2011).
The discrepancy regarding neural correlates of implied motion could be due to differences in experimental methodology (e.g., controlling for low level features, contrast, and timing). It is evident that there are no standardized methods of presenting implied motion in regards to timing and image properties (e.g., color, size, contrast, etc).
For example, the stimuli include, but are not limited to, grayscale images of (in)animate objects (Kourtzi, & Kanwisher, (2000); Jellema & Perrett 2003; Pavan, Cuturi, Maniglia, Casco, & Campana (2011), color images of (in)animate (Winawer, Huk, & Boroditsky, (2008), Fawcett, Hillebrand, & Singh, (2007); Holmes & Wolff, (2010), color images of hand in grasping movement (Urgesi, Moro, Candidi, & Aglioti, (2006), and cartoon images in articulated positions (Osaka, Matsuyoshi, Ikeda, & Osaka, (2010). In addition, the presentation of the stimuli varies as a function of experimenters’ parameter (i.e., motion perception, motor perception, and time perception) from approximately 25ms to
Winawer, Huk, & Boroditsky, (2008), Fawcett, Hillebrand, & Singh, (2007), Urgesi, Moro, Candidi, & Aglioti, (2006). Taken together, the image properties and the rate of presentation may influence different regions of the brain in a top-down process. For example, in a study by Proverbio, Riva, & Zani, (2009), participants were exposed to relatively bright (i.e., 47.88 cd/m2) and detailed colorful images of humans (e.g., athletic women running) presented at a rate of 1500 millisecond (ms) per frame. They found activation in MT, EBA, STS, pre-motor (BA-6), motor areas (BA-4), cingulate, and IF cortex. The method and features of the stimuli may, in part, explain the area of the brain that was activated in conjunction with area MT/MT+. In this paper, the properties of the stimuli and the rate of presentation were strategically chosen to reduce the influence of high-order processing on area MT/MST+.
Interestingly, implied motion is not limited to physical stimuli, but also nonphysical and imaginary stimuli (Saygin, McCullough, Alac, & Emmorey, (2010), Tartaglia, Bamert, Mast, & Herzog (2009); Tartaglia, Bamert, Herzog, & Mast, 2012).
For example, Saygin, McCullough, Alac, & Emmorey, (2010) found that motion sentences (i.e., sentences that include verbs) activated area MT+ at significantly higher rate than static sentences (e.g., sentence that does not describe action). Furthermore, there are two studies (Tartaglia, Bamert, Mast, & Herzog (2009); Tartaglia, Bamert, Herzog, & Mast, 2012) that examined the effects of perceptual training of imaginary stimuli on performance of real stimuli. More specifically, Tartaglia, Bamert, Mast, & Herzog (2009) showed that by instructing participants to imagine the offset of the middle line in a threeline bisection task2, improved performance on a bisection task with three physical lines.
and offset to the right (high frequency tone). Tartaglia, Bamert, Herzog, & Mast, (2012) successfully expanded on the three-line bisection task study through the examination of an imaginary motion discrimination task. Performance on imaginary motion discrimination task led to performance enhancement in real motion discrimination task. It has been empirically shown that multisensory feedback (audio or visual) increase performance in visual perception tasks (Seitz, Kim, & Shams, 2006). Based on pervious literature, multisensory feedback was utilized to enhance performance in implied motion training.
Although there is growing interest in neural correlates of implied motion, no study, to our knowledge, has examined the effects of implied motion, as a visual experience, on low- and high-level cortical processing. The goal of this paper is to expand on Seitz, Nanez, Holloway, & Watanabe’s, (2006) findings, through examination of implied motion training. They found that perceptual learning of directional-motion (i.e., visual experience) led to higher Critical Flicker Fusion Threshold (CFFT) (i.e., alteration of cortical processing). More specifically, functional change in motiondirection accompanied significant change in temporal processing of flicker perception, compared to control, non-coherent motion, and no-motion groups. Therefore, directional motion is an important visual cue that will be utilized in this study.
CFFT refers to a critical frequency of intermittent light (i.e., number of on and off flicker per second) that are perceived as steady continuous light to the human observer.
Previous research has indicated that CFFT is inversely correlated with frontal lobe lesions (Halstead, 1947); mental disorders (Saucer & Sweetbaum, 1958; Curran & Wattis, (2000)