Biology, Neuroscience

The ability to coordinate rhythmic finger movement with a metronome is constrained by both the target timing relation and the rate of coordination. For slow metronome rates (<2 Hz), subjects are able to both syncopate (in between successive beats) and synchronize (on each beat) with the metronome. At faster rates, however, syncopation becomes unstable and subjects spontaneously switch to synchronization in order to maintain a 1:1 stimulus/response relationship. No switches are observed if subjects start in synchronization, indicating it is an inherently more stable mode of coordination. Patterns of brain activity associated with transitions from syncopation to synchronization as well as synchronization only were examined as the metronome rate was increased from 1.0 to 2.75 or 3.0 Hz. Significant differences in the power of the coordination frequency component of event-related potentials (EEG) as well as the MEG beta (15--30 Hz) rhythm were observed when brain activity associated with syncopation was compared to that accompanying synchronization. These differences were focused over left central and centro-parietal areas and the direction of difference in both cases suggests that syncopation is associated with stronger activation of contralateral sensorimotor cortex. Similar results were found when subjects only imagined performing each mode of coordination at an increasing rate, indicating that differences in signal power at least partially reflect neural processes associated with motor planning and preparation independent of overt execution. Consistent with these findings, functional MRI revealed syncopation to be accompanied by significantly more activity in a wide array of cortical (e.g., premotor, prefrontal) and subcortical (basal ganglia, cerebellum) areas known to play a role in motor planning and/or timing of behavior. Whereas the neuromagnetic auditory response decreased as function of coordination rate, the motor evoked response remained approximately constant. This was true both when subjects syncopated and synchronized but may reflect changes in auditory-motor integration near movement rates that induce transitions in the former case. A control experiment examined only self-paced movement and showed a second neuromagnetic motor 'readiness' response that was strongly attenuated for rates above 1.0 Hz. This may signify a decreased need for the planning of motor behavior at faster rhythmic rates.

The survival of rat retinal ganglion cells (RGCs) after axotomy has been shown to be enhanced by Brain Derived Neurotrophic Factor (BDNF). It was, therefore, of interest to determine whether previously observed changes in the differential regulation of fast axonally transported proteins (FTPs) occur in rat RGCs during the early response to axotomy or whether such changes are obviated by the action of BDNF at the cell body level. It was of further interest to determine whether these regeneration-associated changes are sustained during the period of BDNF-enhanced cell survival. It was found that, within 2 days of injury and BDNF injection, rat RGCs initiate a growth-like cellular response that includes the differential synthesis and transport of the same profile of FTPs found to be induced in axotomized animals following injection of a saline control solution. Thus, supplementation of rat RGCs with BDNF does not obviate the changes required to reinstate active cellular regrowth. It is, therefore, unlikely that the loss of a trophic factor, such as BDNF, is the signal for axotomy-induced changes. Although a single injection of BDNF at the time of injury prolongs cell survival to at least 5 days, it is not sufficient to sustain the elevation in FTPs. This result indicates that the regulatory mechanisms that promote cell growth are distinct and separate from those that promote cell survival. This study extended beyond the above findings to affirm that apoptosis of axotomized rat RGCs is mediated by the activation of the cysteine protease, caspase-3. Such activation was demonstrated within 12 hours of axotomy and appeared to become increasingly prevalent in a central to peripheral gradient, as might be anticipated by the loss of glial derived neurotrophic support. Such activation was completely prevented by intraocular injection of BDNF, indicating that BDNF acts upstream of caspase-3 to prevent the proteolytic cascade that leads to apoptosis.

Mismatch Negativity (MMN) is a component of the event-related potential (ERP) that is associated with the detection of novel stimuli in one's environment. Naatanen has suggested that a neural template theory provides the best explanation of the mechanism that underlies this psychophysiological phenomenon. The purpose of the present project was to propose that a Hebbian model of cell-assemblies (Hebb, 1949) provides a plausible competing theory of MMN. A Hebbian model is consistent with the evidence provided by imaging studies that demonstrate increased neural efficiency in learning cognitive (as opposed to skilled motor) tasks and with recent animal studies in an analogous model. This model suggests three hypotheses which were addressed by the present study. First, it is proposed that the method that is traditionally used to calculate MMN may not be the ideal. Specifically, it is proposed that the baseline measure used in the calculation does not yield optimal MMN data and the present experiment investigated a new method of gathering baseline data. Second, it was hypothesized that an investigation of sequence effects related to standard and deviant stimuli in an oddball paradigm would provide further support for a Hebbian reinterpretation of MMN. Finally, the argument is made that a Hebbian model of MMN garners additional support in terms of parsimony and ecological validity in addition to being consistent with our current understanding of the physiological underpinnings of learning phenomena. Results indicate that the new method of calculating MMN does, indeed, provide a more robust measure of MMN. Furthermore, sequence effects were demonstrated for both the standard and deviant stimuli, however the sequence effects observed in standard stimuli were not in the expected direction. Both practical and theoretical implications are discussed.

Following Kelso et al. (1991a; 1992), Wallenstein et al. (1995), Tuller et al. (1994), and Case et al. (1995); see also Fuchs et al. (1992), the experiments described in this research all used a dynamical methodology designed to produce a coherent brain state and then lead that brain state through a spontaneous reorganization via the influence of a parametric change. Magnetoencephalographic (MEG) recordings made during the spontaneous behavioral and perceptual transitions were analyzed by decomposition of the brain's high-dimensional magnetic field into a few task-relevant components. The analysis showed that the dynamics of the MEG signal, including the reorganization which occured as a result of the parametric manipulation, could be accounted for by the dynamics of the individual components. This supports the idea that the task requirements in each case placed the brain into a (relatively) low-dimensional state through the cooperative interactions among the many neuronal elements involved in the task. The experiments included two coordination experiments in which subjects were required to produce index finger flexions in time to an auditory metronome in an anti-phase pattern while the metronome rate was increased. Increases in the variability of both the behavior and the motor-associated magnetic field components prior to the transition to an in-phase pattern support the hypothesis that a dynamic instability mechanism exists for pattern formation and change during those tasks. In the third experiment a perceptual instability was explored by systematically scaling a parameter known to influence categorization of speech stimuli: biasing the transition created stimuli that were perceived in two different ways. The design of the experiment allowed the investigation of neural correlates of the physical properties of the stimuli, perceptual invariance, bistability, and perceptual reorganization. Analysis of the MEG signals suggests that presentation of a bistable stimulus places the brain into a highly sensitive, unstable state that can be influenced by ongoing activity.

Recent studies have demonstrated that the strategy an individual uses to synchronize motor behavior (e.g. finger flexions) with externally delivered, periodic stimuli depends, in part, on the stimulus presentation rate (Mates, Muller, Radil, and Poppel, 1994; Engstrom, Kelso, and Holroyd, 1995). At rates slower than approximately 0.5 Hz, subjects typically exhibit a reactive-type coordination pattern where the response follows the stimulus by an order of magnitude consistent with typical response times (i.e. 150-250 milliseconds). At faster rates, however, subjects typically anticipate the impending stimulus in order to synchronize movement with it. In the present study, scalp electroencephalographic (EEG) signals (61 channels) were recorded during a sensorimotor task designed to investigate transitions from one coordination mode to another. We found that subjects exhibited a spontaneous transition from reactive to anticipatory behavior as the stimulus presentation rate increased past some critical frequency. A spatiotemporal analysis of the EEG signals accompanying this task revealed: (1) a widespread frequency component in the EEG matching that of both the stimulus and movement; (2) peak spectral power density over central and antero-central regions in both men and women during reactive behavior; (3) an additional bilaterally distributed frontal component at the most anterior portion of the scalp in men during anticipatory behavior; (4) an additional left fronto-central component which extended posteriorly toward antero-central regions in women during anticipatory behavior; (5) fluctuation enhancement in both the EEG spectral power density and the time lag ($\tau$) between the movement and stimulus accompanying the transition from reactive to anticipatory behavior; (6) that the spectral power density patterns obtained in the primary experimental condition (REACTIVE) were more similar in terms of their spatial distribution with a control condition in which subjects were asked to produce rhythmic movements without benefit of an external stimulus (MOTOR-ONLY) than with a control condition in which subjects passively watched a periodic visual stimulus (STIMULUS-ONLY); (7) that the spectral power density patterns obtained during reactive behavior in the primary experimental condition were more similar spatially to the MOTOR-ONLY condition than du ring anticipatory behavior; and finally (8) that the spectral power density patterns obtained during the experimental condition are not completely accounted for in terms of purely motor- or stimulus-related components. These results are discussed within a common framework of pattern formation instigated by dynamic instabilities in the human brain and behavior.

Available evidence suggests that the median raphe nucleus (MRN), when activated, produces a desynchronized hippocampal electroencephalog ram (EEG), and that this effect is sensitive to serotonergic (5-HT) manipulations. Experiment 1 examined the effect of injections into the MRN of agents that non-specifically (procaine) or selectively (8-OH-DPAT and buspirone) inhibit serotonin-containing MRN neurons. These substances produced hippocampal theta rhythm at short latencies and for long durations, suggesting that MRN 5-HT neurons are specifically responsible for controlling the hippocampal EEG. MRN 5-HT neurons are modulated by a facilitatory excitatory amino acid (EAA) input and an inhibitory influence from GABAergic interneurons within the MRN. Experiments 2 and 3 examined the effect of manipulations of these systems on the hippocampal EEG. Experiment 2 demonstrated that injections of the specific (AP-7) and non-specific (MK-801) NMDA antagonists, as well as the kainate/quisqualate antagonist (GAMS) into the MRN produce theta at short latencies and for long durations. Experiment 3 demonstrated that injections of the GABA$\sb{\rm A}$ agonist, muscimol, into the MRN produced hippocampal theta rhythm at short latencies and for long durations. In light of recent evidence suggesting a theta-pacemaker role for numerous brain nuclei, experiment 4 sought to re-examine the role of the medial septum/diagonal band complex (MS/DB) in hippocampal theta rhythm produced by injections of 8-OH-DPAT into the MRN. Four categories of MS/DB neurons were described: (1) cells which burst rhythmically with theta (rhythmical); (2) cells displaying a tonic increase in discharge with theta (theta-on); (3) cells displaying a dramatic decrease or cessation of discharge with theta (theta-off); and (4) cells which showed no changes in discharge in relation to theta (no-change). It was shown that injections of 8-OH-DPAT into the MRN caused a change in discharge of rhythmic MS/DB cells from an irregular non-bursting pattern during baseline conditions to a rhythmical, bursting pattern which was highly coherent with the hippocampal EEG.

Taurine is one of the most abundant amino acids in mammals and several functions of taurine have been reported. One important function of taurine is its neuroprotection against the glutamate-induced neuronal damage. It was shown that the glutamate-induced neurotoxicity is caused by overexcitation of glutamate receptors and intracellular calcium, [Ca2+]i, elevation. In this dissertation, the mechanism underlying the action of taurine as a neuroprotector was investigated. It was found that taurine protected neurons against glutamate or Bay K 8644-induced neurotoxicity only at the concentration that inhibits the calcium influx induced by those two compounds. Furthermore, taurine couldn't protect neurons against sodium nitroprusside, a NO free radical donor, induced neurotoxicity. These results indicate that taurine exerts its neuroprotection by reducing the glutamate-induced [Ca2+]i elevation. Besides necrosis, apoptosis is another major way that glutamate induces neuronal cell death. The effect of taurine on the glutamate-induced apoptosis was investigated. It was found that taurine prevented the glutamate-induced DNA fragmentation, indicating taurine prevents the glutamate-induced apoptosis. We found that anti-apoptotic proteins (BCL-2 and BCL-X) were down-regulated by glutamate treatment and this down-regulation was prevented by taurine. No difference in pro-apoptotic proteins (BAX and BAD) was found. It was found that the down-regulation of BCL-2 and BCL-X was through calpain-mediated proteolysis, and taurine may exert its anti-apoptotic function by preventing the activation of calpain, which is due to the prevention of [Ca2+]i elevation. Furthermore, it was found that pre-treatment with taurine inhibited the glutamate-induced calcium influx through L-, P/Q-, N-type voltage-gated calcium channels and NMDA receptor. Surprisingly, taurine had no effect on calcium influx through the NMDA receptor when neurons were treated with NMDA in Mg 2+-free medium. The effect of taurine is unlikely through GABA A, or glycine receptors, since bicuculline and picrotoxin (GABA A receptor antagonists), and strychnine (glycine receptor antagonist), failed to block taurine's inhibitory effect on the glutamate-induced calcium influx. Since taurine was found to prevent the glutamate-induced membrane depolarization, we propose that taurine protects neurons against the glutamate excitotoxicity by preventing the glutamate-induced membrane depolarization, probably through the opening of chloride channels, therefore preventing the glutamate-induced calcium influx and the downstream events.

In the central nervous system (CNS), the rate-limiting step in GABA synthesis is the reaction catalyzed by L-glutamic acid decarboxylase (GAD). Alternations in the level of GABA or GAD have been linked to various neurological disorders. Mammalian species express two isoforms of GAD, namely, GAD65 and GAD67, referring to GAD with a molecular weight of 65 kDa and 67 kDa, respectively. Numerous studies have been done to elucidate the mechanisms that control the regulation of GAD at the level of gene expression, protein synthesis, saturation of co-factor, pyridoxal 5'-phosphate (PLP), and post-translational modification. Our previous studies had demonstrated the presence of the truncated form of human brain L-glutamic decarboxylase 65 (tGAD65) in vivo as well as in vitro and found that tGAD65 was more active than the full-length GAD65 (fGAD65). In addition, the recombinant human brain GAD67 has been found to be specifically cleaved at two specific sites, one at arginine 70 and another at arginine 90, to produce two truncated forms of GAD 67 (tGAD67). It seems that the formation of tGAD is catalyzed by specific proteases instead of a random degradation. Furthermore, it has been found that GAD65 is regulated by the Ca2+-free form of calmodulin (apoCaM). My research focus is to elucidate the regulation of GABA biosynthesis through regulation of its synthesizing enzyme, especially GAD67, by protein phosphorylation, proteolytic cleavage and apoCaM. Experiments presented here have been conducted to demonstrate the molecular cloning, expression, and purification of human brain tGAD67. The purified protein was further characterized by kinetic studies and phosphorylation studies. Truncated forms of hGAD67 were much less active than the full-length form. Both truncated enzymes are also phosphorylated by protein kinase A (PKA) as is full-length hGAD67. A deletion of 1-70 aa from the N-terminal results in additional protein kinase C (PKC) phosphorylation. Several phosphopeptides and possible phosphorylation sites are suggested by matrix assisted laser desorption/ionization-time of flight (MALDI-TOF) analysis. Furthermore, evidence of mu-calpain, not m-calpain, as the protease responsible for GAD cleavage in vivo as well as in vitro is presented. In addition, evidence on the effect of ApoCaM on GAD67 activity, phosphporylation and proteolytic cleavage by mu-calpain is discussed. Finally, an overall model of GAD regulation by a variety of mechanisms including protein phosphorylation, mu-calpain proteolytic cleavage and apoCaM is proposed.

In this dissertation, we examined the neural correlates of motor coordination and music perception using a set of four fMRI experiments. The neural correlates of goal-directed action were examined in a group of healthy adults in experiment I using execution and imagery of a unimanual and a bimanual finger-sequencing task. Similar neural networks were engaged for execution and imagination of movement sequences. Interestingly, we also found that the sensorimotor cortical and cerebellar areas are functionally decoupled from the task network when people imagine but do not actually execute sequential actions. In experiment 2, we used the same finger-sequencing paradigm to study recovery of function during recovery from stroke. It was observed that the wide spread neural activity during the initial session became more localized during the last session. In addition, using imagery tasks, we showed that hemiplegic patients retained the ability to activate neural pathways that are normally involved in executing goal-directed action sequences, despite the loss of ability to actually execute movements. In experiment 3, we examined brain activity when musicians and non-musicians listened to expressive and mechanical versions of a musical piece. The expressive performance activated the limbic areas more than the mechanical version in both groups of subjects suggesting perception of affect. The pattern of neural activity was also dictated by their experience and familiarity with the piece of music. In addition, we found activation of language related areas when musicians listened to the expressive version suggesting shared neural resources for language and music. The neural basis of sensorimotor coordination and timing in Parkinson's disease was investigated in the last experiment, using a synchronization-syncopation task and the continuation paradigm. Different neural areas subserved timing during the two different modes of coordination. However, these differences persisted during their respective continuation phases. In order to compensate for the functional deficiency in Parkinson's disease, patients recruited functionally segregated circuits that connect the striatum and association areas of the parietal, premotor and prefrontal cortices.

A prominent deficit in Alzheimer's disease (AD) is a difficulty in recognizing and naming people. Unfamiliar and famous face recognition tasks are sparse in the neuropsychology literature. It was hypothesized that: a deficit in recognition of faces would be found for AD patients, the semantic mismatch condition would result in the longest response latency and least accurate naming, and semantic cues would not facilitate naming for the AD group. Accuracy and reaction time from ten mild AD patients diagnosed by NINCDS-ADRDA criteria and 10 normal controls (matched age, 66--82 years, education & ethnicity) were tested via unfamiliar and famous faces recognition memory tests and famous faces naming tasks with and without semantic interference. Both subject groups were more accurate on the famous face recognition rather than memory for unfamiliar faces, with significant group differences. The Stroop-like face naming task performance was characterized by an increased interference effect, semantic face-name mismatches produced the longest response delays, and less accurate face naming particularly in the AD group. The semantic cues resulted in a decrease in naming accuracy for the AD patients, which may be indicative of their name retrieval deficit. Consistent with existing face-name models, these findings suggest that the deficit in AD is related to semantic naming rather than the perceptual component of face recognition. Furthermore, the ability to correctly name faces even in the presence of interference may prove to be a diagnostic tool that is sensitive to face naming deficits characteristic in cases of brain damage.