1摘要

1.1課題1
在體內的新生大鼠桶狀皮層長時程增強狀態
“長時程增強（LTP）的活性依賴的形成的早期皮質電路是很重要的。在新生鼠桶狀皮層LTP已到目前為止僅在體外研究。我結合電壓敏感染料成像與胞外多電極記錄研究兩個斜坡的場電位和對新生大鼠桶狀皮層在體內的桶狀皮層通路在晶須的多單元活動人數晶須刺激誘導的LTP。10分鐘，在2赫茲單晶須刺激誘導的LTP在出生后的年齡依賴性的表達（P）0～P14大鼠在p3-p5 LTP的表達最強。LTP的幅度在刺激筒柱最大相關，較小的周邊間隔區和沒有LTP可以在相鄰的桶的觀察。電流源密度分析顯示相關的LTP增加突觸電流沉在第四層/低層II / III在p3-p5在皮質板/上層V在p0-p1。這項研究表明，為第一次在體內的新生大鼠桶狀皮層LTP的年齡依賴性和空間限制。這些活性依賴的修改中的關鍵時期，可能與在每桶皮質地形圖的細化的發展起著重要的作用。”

1.2課題2
在新生大鼠皮層通過伽馬和主軸陣陣觸發的早期活動

主軸陣陣觀察M1可能引發早期的爪子運動。此外，爪運動還可以引起M1層V.局部失活的M1可以明顯降低爪運動的局部電刺激，這表明新生兒M1在電機模式。相反，新生兒M1也感覺模式操作。早期的自發運動和M1爪觸發γ和主軸陣陣感官刺激。阻斷外周感覺輸入從爪完全廢除誘發γ和主軸陣陣。此外，感覺誘發和自發產生的γ和主軸陣陣介導的S1和M1之間的相互作用。因此，在S1深刻地降低爪刺激誘導的局部滅活和自發產生的γ和M1主軸陣陣，表明S1的活動模式，在M1代起著至關重要的作用。本研究提出，自我生成和感覺誘發γ和M1主軸陣陣可能有助于改進和皮質脊髓束和感覺在感覺運動協調需網絡成熟。

1 Summary
1.1 Project 1
Long-term potentiation in the neonatal rat barrel cortex in vivo

“Long-term potentiation (LTP) is important for the activity-dependent formation of early cortical circuits. In the neonatal rodent barrel cortex LTP has been so far only studied in vitro. I combined voltage-sensitive dye imaging with extracellular multi-electrode recordings to study whisker stimulation-induced LTP for both the slope of field potential and the number of multi-unit activity in the whisker-to-barrel cortex pathway of the neonatal rat barrel cortex in vivo. Single whisker stimulation at 2 Hz for 10 min induced an age-dependent expression of LTP in postnatal day (P) 0 to P14 rats with the strongest expression of LTP at P3-P5. The magnitude of LTP was largest in the stimulated barrel-related column, smaller in the surrounding septal region and no LTP could be observed in the neighboring barrel. Current source density analyses revealed an LTP-associated increase of synaptic current sinks in layer IV / lower layer II/III at P3-P5 and in the cortical plate / upper layer V at P0-P1. This study demonstrates for the first time an age-dependent and spatially confined LTP in the barrel cortex of the newborn rat in vivo. These activity-dependent modifications during the critical period may play an important role in the development and refinement of the topographic map in the barrel cortex.” (An et al., 2012)#p#分頁標題#e#


1.2 Project 2
Early motor activity triggered by gamma and spindle bursts in neonatal rat motor cortex
Self-generated neuronal activity generated in subcortical regions drives early spontaneous motor activity, which is a hallmark of the developing sensorimotor system. However, the neuronal activity patterns and functions of neonatal primary motor cortex (M1) in the early movements are still unknown. I combined voltage-sensitive dye imaging with simultaneous extracellular multi-electrode recordings in the neonatal rat S1 and M1 in vivo. At P3-P5, gamma and spindle bursts observed in M1 could trigger early paw movements. Furthermore, the paw movements could be also elicited by the focal electrical stimulation of M1 at layer V. Local inactivation of M1 could significantly attenuate paw movements, suggesting that the neonatal M1 operates in motor mode. In contrast, the neonatal M1 can also operate in sensory mode. Early spontaneous movements and sensory stimulations of paw trigger gamma and spindle bursts in M1. Blockade of peripheral sensory input from the paw completely abolished sensory evoked gamma and spindle bursts. Moreover, both sensory evoked and spontaneously occurring gamma and spindle bursts mediated interactions between S1 and M1. Accordingly, local inactivation of the S1 profoundly reduced paw stimulation-induced and spontaneously occurring gamma and spindle bursts in M1, indicating that S1 plays a critical role in generation of the activity patterns in M1. This study proposes that both self-generated and sensory evoked gamma and spindle bursts in M1 may contribute to the refinement and maturation of corticospinal and sensorimotor networks required for sensorimotor coordination.

2 Introduction引言

2.1 Project 1
2.1.1 Sensory pathways from whiskers to barrel cortex
 感覺通路來自胡須桶狀皮層
The lemniscal and paralemniscal pathways transfer the sensory information from the whiskers to the barrel cortex (Fig. 1). In the brain stem, the lemniscal pathway begins at the principal trigeminal (PrV) nucleus and sends fibers to the ventral posteromedial (VPM) nucleus in the thalamus. While the paralemniscal pathway starts in the spinal trigeminal nucleus (SpVi) of the brain stem and projects to the posterior (POm) nucleus in the thalamus. Then, the lemniscal pathway projects the thalamoscortical fibers from VPM to the layer IV barrels in the barrel cortex. Comparatively, the paralemniscal pathway targets the layer IV septa through the thalamocortical projections from POm ( for review Lubke and Feldmeyer, 2007;Alloway, 2008).
The lemniscal pathway has a precise topographic organization and each whisker has a well-defined columnar organization. The PrV and VPm nuclei contain “barrelettes” and “barreloids”, respectively, which correspond to the spatial distribution of the cortical barrels. Those organizations can maintain the spatial relationships for the neighbouring barrels (Vanderloos, 1976;Jacquin and Rhoades, 1983;Killackey and Fleming, 1985). In the newborn rat, a single whisker deflection activates the corresponding cortical barrel and produces weak neuronal responses in the neighboring septal regions (Yang et al., 2012). The lemniscal system has the specific single-whisker distribution. It can encode spatial and temporal information, such as identification of orientation, shape, distance and other spatial features (Krupa et al., 2001;Harvey et al., 2001;Shuler et al., 2002). However, the paralemniscal pathway has multi-whisker distribution. Neurons in POm and in the cortical septa receive multiple whiskers inputs (Diamond et al., 1992). Furthermore, the paralemniscal system is much poorer in encoding the spatial and temporal encoding information, but more effective in encoding the rate of passive whisker movements than those in the lemniscal system (Melzer et al., 2006a;Melzer et al., 2006b).#p#分頁標題#e#

Hence, the lemniscal pathway processes spatiotemporal information by whisker contact with external objects, while the paralemniscal pathway encodes the frequency and other kinetic features of active whisker movements. Both pathways work cooperatively with each other to identify the external environment by passive or active whisker movements.

information in to the nucleus in the brain stem. One stream goes through the spinal nucleus, then it arrives at the VPM (pink), while the other goes through the principal nucleus, then reaches at the PoM (light blue). Finally, thalamic afferents arising either from neurons in the VPM (red line) or POm (green line) project to different cortical laminae in the somatosensory barrel field (framed area) of the neocortex (modified from Lubke and Feldmeyer, 2007). B. Schematic diagram illustrating the lemniscal (red) and paralemniscal pathway (green) from brain stem to the barrel cortex. C. The left diagram, photomicrograph of tangential section through layer IV of the barrel field after processing the tissue by cytochrome oxidase staining. The right diagram, which is based on the photomicrograph, designates each barrel according to its arc position (1-5) within a specific row (A-E). Although many septal zones are no more than 70-80 μm in width, the septal regions indicated by the asterisks are more than 200μm wide, R, rostral; L, lateral (modified from Alloway, 2008).

2.1.2 The development of sensory pathways in the barrel cortex
發展感覺通路在每桶皮質
During the last embryonic and first postnatal week, the rat barrel cortex undergoes dramatic developments (Fig. 2). As early as E13, thalamocortical axons and early corticofugal axons arrive in the ventral region of the telencephalon synchronously and are thought to interact. By E17, both lemniscal and paralemniscal pathways have crossed the pallial-subpallial border (PSPB) and have reached the cortex, where they extend tangentially in the intermediate zone (IZ). At the same time, the development of the laminar structure begins, the marginal zone (MZ) and subplate (SP) generate from the cortical plate (CP). Immediately after birth at P0, both lemniscal and paralemniscal pathways have already arrived at the CP and layer V, respectively. Two days later, barrel patterns generate in layer IV by the axons from the lemniscal pathways. At P7, these patterns become more mature. Surprisingly, the paralemniscal pathways reach the superficial layer MZ through layer II/III (for review Price et al., 2006; Galazo et al., 2008).

formation which appears by 24 hours after birth. Finally, the barrel-related patches pattern became more mature at 72 hours. Therefore, these suggest that the thalamic afferents show a somatotopic pattern in cortex as early as the time of birth.

Fig. 3 Anatomical study shows the development of barrels in neonatal rats.
Tangential distribution of AChE-reactive afferents from the flattened cortices of four littermates. A, As early as 4 hours after birth, there are five separate rows (AChE-reactive afferents) can be observed in the tangential sections. B, One hour later, the five rows are more clearly separated (represented by a-e). C-D, the jaw (j), forelimb (f) and hindlimb (hl)representations are less clear than the whisker representation during the 1st postnatal day. E-F, after about 38 hours, clear and precise isomorphic map of whiskers in barrel cortex can be identified. r, rhinal vibrissae Scale bar = 675 μm (modified from Schlaggar and Oleary, 1994).#p#分頁標題#e#

2003). However, more and more scientists found that long-term potentiation (LTP) is important for activity-dependent neuronal development in rodent neocortex (Feldman et al., 1999;Daw et al., 2007;Inan and Crair, 2007).

2.1.4 LTP at thalamocortical synapses
LTP在丘腦皮層突觸
Previous studies suggested that LTP is a likely candidate as the synaptic mechanism underlying certain forms of learning and memory (Morris, 2003). However, there is another considerable evidence showed that the NMDA-dependent LTP serves for the correct formation and refinement of receptive fields in the barrel cortex (Crair and Malenka, 1995;Daw et al., 2007). Chronic application of the NMDA receptor antagonist (AP5) in the barrel cortex during the first postnatal week could disturb the precise topographical map and exhibit deficient experience-dependent receptive field plasticity (Schlaggar et al., 1993;Fox et al., 1996). Furthermore, in neonatal rat barrel cortex, LTP can also be induced in vitro by a pairing protocol from postnatal day (P) 3 to 7 (Fig. 4) (Crair and Malenka, 1995;Isaac et al., 1997;Barth and Malenka, 2001), the time period that coincides with barrel map formation (Fox, 1992;Fox et al., 1996;Foeller and Feldman, 2004). The magnitude of LTP in the barrel cortex gradually decreases between P3 and P8 (Crair and Malenka, 1995). Similarly, synaptic plasticity (ocular dominance plasticity, ODP and LTP) is important for the development of the visual cortex (Malenka and Bear, 2004). For example, a long time period in the dark for animals could result in a longer critical period for ODP and LTP in the visual cortex (Kirkwood et al., 1995). Tetrodotoxin application in the open eye abolished the ODP (Antonini and Stryker, 1993). Again postsynaptic NMDA receptor antagonists also abolish ODP and LTP (Bear et al., 1990;Daw et al., 1999). All these studies suggest that NMDA receptor-dependent synaptic plasticity is involved in the formation and refinement of receptive fields in the primary sensory systems.

P2-P3, about 30% of thalamocortical synapses keep “silent”, based on comparison of failure rates at hyperpolarized and depolarized potentials (Fig. 5). Interestingly, in P4-P5 animals, the proportion of silent synapses increases to 40%. However, after P5, the proportion of silent synapses decreases dramatically. At P8-P9, there is no silent synapses at all (Fig. 5) (Isaac et al., 1997). These observations are in good agreement with previous reports on the age-dependent expression of LTP in thalamocortical slice preparation of the newborn rat (Crair and Malenka, 1995). As show in Fig. 4e, the strongest LTP can be induced at P3-P5 indicating that the large amount of the silent synapses reveals the highest capability for activity-dependent change. After P5, the proportion of silent synapses falls sharply which may result in the weak LTP during these ages. The pairing stimulation may cause the rapid appearance of AMPA currents which converts silent synapses to functional ones. All above observations suggest that thalamocortical synapses are initially born with silent (post)synaptic elements which become functional by LTP (Feldman et al., 1999).#p#分頁標題#e#

Fig. 5 Silent thalamocortical synapses.
A.Voltage-clamp recording in a layer IV cell. Note no AMPAR-mediated EPSCs holding at -70 mV, but robust EPSCs exist when depolarizing to +50 mV. D-APV (25 μM) blocked the EPSCs observed at +50 mV. Each point represents the amplitude of an individual EPSC. B. Average of 40 responses (top) and eight superimposed consecutive responses (bottom) for the following epochs of the experiment shown in A, respectively. C. Bar diagram of success rates at depolarized versus hyperpolarized potentials for thalamocortical EPSCs in minimal stimulation experiments at different ages. Note that the silent synapses exist during the critical period. Data are expressed as mean ± s.e.m. (modified from Isaac et al., 1997).

2.2 Project 2
2.2.1 Central pattern generators (CPGs) generate early motor activity
中樞模式發生器（CPGS）產生的早期活動
Brainstem activity in young infants and during active sleep were firstly suggested to induce ascending activation of the cortex and descending activation of the musculature (Fig. 6) (Roffwarg et al., 1966). The hypothesis has been elaborated that brainstem-initiated twitching of the paws generates sensory feedback and then activates the cortex (for review, see Blumberg, 2010b). However, in mammals, including humans, the spinal cord has been suggested as CPG for locomotion (MacKay-Lyons, 2002;Dietz, 2003;Ijspeert, 2008;Guertin, 2009;Minlebaev et al., 2011). In sleeping newborn rats, spontaneous muscle twitches send the tactile feedback to the spinal cord and guide the organization of spinal sensorimotor circuits (Petersson et al., 2003). In adult animals, forelimb representation in M1 has been suggested to pertain to three different categories of movement, for example, stereotyped repetitive behaviors, complex voluntary movements and fine motor manipulation skills (for review, see Levine et al., 2012).

Fig. 6 Spinal cord and brainstem generate movements.
A.Schematic illustrations of organization of spinalcord and paw reflex module. The primary sensory afferents send the tactile information through monosynaptical and oligosynaptical circuits. Then the 2004). Central pattern generators (CPGs) synaptic connections have been suggested to be critical for these behaviors (Clarac et al., 2004). Moreover, different electromyography (EMG) patterns and duration were observed in various behaviors, i.e. characteristic asymmetric EMG waveform with double bursts contributes to the stepping; the bell-shaped EMG curve with only single burst is necessary for the swimming (Gruner and Altman, 1980;Deleon et al., 1994).

Fig. 7 The development of the locomotion behaviors in the rat.
The development of locomotion can be divivded into fetal, immature, transitory and adult stages. Motor neurons and central pattern generators have already operated during the fetal stage. The immature stage starts at P0 and ends at P10. During this period, the corticospinal tract has already connected the motor cortex to the spinal cord. Meanwhile, some basic and simple behaviors can be oberserved. From P10 to P15 is the period for the transitory stage. During this stage, the rat eye and ear are opening, meanwhile they can perform whisking. Fanally, in the adult stage, they can perform adult swimming and walking with high motivation (modified from Clarac et al., 2004).#p#分頁標題#e#

2.2.3 Early neuronal activity patterns triggered by sensory feedback modify the functional topography map
早期的神經活動模式引發的感覺反饋修改功能地形圖
Although genetically determined molecular factors considerably contribute to functional sensorimotor connectivity (for review, see Polleux, 2005), early synchronized oscillatory network activities are also believed to be essential for the formation of the functional neuronal circuits (for review, see Khazipov and Luhmann, 2006) and maintenance of tonotopic maps in neonatal cortex (Tritsch et al., 2007;Colonnese et al., 2010;Minlebaev et al., 2011;Yang et al., 2012). These early activity patterns control axon growth and synaptogenesis (for review, see Allene and Cossart, 2010). As show in Fig. 8, sensory periphery input could also trigger early activity patterns, for example, early movements drive spindle bursts in the neonatal primary somatosensory cortex (S1) (Khazipov et al., 2004). In the early whisker system, spontaneous whisker twitches (Tiriac et al., 2012) and whisker sensory stimulation elicit both gamma and spindle bursts which synchronize developing thalamus and barrel cortex (Yang et al., 2009;Minlebaev et al., 2011;Yang et al., 2012). In visual system, retinal waves trigger spindle bursts in the neonatal visual cortex (Hanganu et al., 2006;Colonnese et al., 2010). In the developing auditory system, supporting cells in the cochlea spontaneously release ATP and process the bursts of action potentials in auditory nerve fibres before the onset of hearing (Tritsch et al., 2007;Tritsch and Bergles, 2010). Conversely, these early patterns of electrical activity drive motor output, e.g. the self-generated bursts in spinal cord (Gramsbergen et al., 1970;Petersson et al., 2003;Clarac et al., 2004) and brain stem (Blumberg, 2010b;Tiriac et al., 2012) triggers spontaneous movements. Even muscles can twitch spontaneously during sleep (Petersson et al., 2003;Mcvea et al., 2012).

molecular mechanisms contribute to the specification and development of corticospinal motor neurons. e.g., a small number of corticospinal motor neurons genes (such as, Diap3, Igfbp4 and Crim1) seem to be restricted to the sensorimotor cortex. Other genes are expressed across the full extent of layer V (e.g., Ctip2, encephalopsin, Fezf2, Clim1, Pcp4 and S100a10) (for review, see Molyneaux et al., 2007). All these genes have been proven to be important for the development of the corticospinal tract. In the absence of the Ctip2 gene, defects in fasciculation, outgrowth and pathfinding were observed in subcerebral projection neuron axons. At the same time, dramatically less axons reach the brainstem. Furthermore, reduction of Ctip2 expression in Ctip2-heterozygous mice results in abnormal developmental pruning of corticospinal axons. All these results suggest that Ctip2 is a crucial regulator of subcerebral axon extension and of the refinement of collaterals as these neurons mature (Arlotta et al., 2005). In addition, after knocking out the Fezf2 gen in null mutant mice, neither subcerebral projection neurons nor corticospinal tract to the spinal cord or the brainstem could be found (Molyneaux et al., 2005). However, it is still unknown that the patterns of activity expressed in primary motor cortex (M1) and whether these activities can transfers through cortical spinal tract during the early developmental stages.#p#分頁標題#e#

Fig. 9 Corticospinal tracts in P0 mice.

A. Schematic representations of sagittal views of the brain and proximal spinal cord in a P0 mice, cortical spinal motor neuron somas in the cortex (red triangles) and their axonal projections toward the spinal cord (red lines). B-D. Representative parts of positions are shown at an expanded scale (right). Note the corticospinal tract stained by Dil from cortex to the spinal cord. E. Schematic diagram of a sagittal section of a mouse brain. Note that the corticospinal motor neurons (purple dot) located at layer V of the cortex project to the spinal cord (modify from Arlotta et al., 2005;Molyneaux et al., 2007).
2.2.5 The development of corpus callosum projections
In newborn rodents, the callosal projection neurons have already crossed the corpus callosum (CC) (Fig. 10) (Tritsch et al., 2007;Molyneaux et al., 2007;Rouaux and Arlotta, 2012). Genetically determined molecular factors control the specification and development of callosal neurons. The Lmo4 gene expressed in callosal neurons of layers II/III and V has been suggested to be important for the development of callosal projection (Arlotta et al., 2005;Molyneaux et al., 2007). However, the neural activity patterns have also been demonstrated to be critical for the development of corpus callosum circuits. In neonatal S1, the weak evoked responses in the ipsilateral hemisphere were observed with EEG (Marcano-Reik and Blumberg, 2008) and VSDI (Mcvea et al., 2012) recordings. Moreover, the weak ipsilateral On-response could be also elicited in V1 by optic nerve stimulation from P3 onwards (Hanganu et al., 2006). Surprisingly, in the whisker system, the whisker stimulation failed to evoke a detectable response in the ipsilateral barrel cortex in the same age of newborn rats (Yang et al., 2009). Nevertheless, in adult mice, the tactile whisker stimulation could induce the reliable VSDI evoked response in bilateral sensorimotor cortices (Ferezou et al., 2007).

Fig. 10 The development of the callosal projection neurons.
A.Schematic representations of a coronal section of the brain in newborn mice. Note the injection position of the DiI crystals (red dots). B. At P0, rare pioneering axons are seen (arrow), prior to entry into the contralateral striatum. C. Increase in the number of axons in the contralateral corpus callosum (arrows) at P2. D-J. At P3-P4, the number of axons in the contralateral corpus callosum increase dramatically. K. Callosal projection neurons. Callosal neurons located at layers II/III, V and VI extend an axon across the CC to the other hemisphere. There are three kinds of callosal neuron: (1) the callosal neurons have single projections to the contralateral cortex (black); (2) the callosal neurons have triple projections to the contralateral cortex and bilateral striatums (blue); (3) the callosal neurons have double projections to the contralateral cortex and ipsilateral frontal cortex (green) (modified from Molyneaux et al., 2007;Sohur et al., 2012

Fig. 10 The development of the callosal projection neurons.
A. Schematic representations of a coronal section of the brain in newborn mice. Note the injection position of the DiI crystals (red dots). B. At P0, rare pioneering axons are seen (arrow), prior to entry into the contralateral striatum. C. Increase in the number of axons in the contralateral corpus callosum (arrows) at P2. D-J. At P3-P4, the number of axons in the contralateral corpus callosum increase dramatically. K. Callosal projection neurons. Callosal neurons located at layers II/III, V and VI extend an axon across the CC to the other hemisphere. There are three kinds of callosal neuron: (1) the callosal neurons have single projections to the contralateral cortex (black); (2) the callosal neurons have triple projections to the contralateral cortex and bilateral striatums (blue); (3) the callosal neurons have double projections to the contralateral cortex and ipsilateral frontal cortex (green) (modified from Molyneaux et al., 2007;Sohur et al., 2012).
Aims of the thesis of project 1

2.11 Aims of the thesis本文的目的是
Whether LTP can be induced by whisker stimulation in the adult barrel cortex in vivo is a matter of debate. Multi-whisker stimulation by air puffs at 5 Hz for 30 s did not induce LTP in adolescent mice (Takata et al., 2011), while multi-whisker stimulation at 2 or 8 Hz for 10 min induced a stable LTP in layers II/III and IV of the barrel cortex of mature mice (Megevand et al., 2009). Whether LTP can be elicited in the neocortex in vivo before the onset of the critical period has not been investigated yet. Therefore, we asked whether single-whisker stimulation may elicit LTP in the neonatal (P0-P7) rat barrel cortex in vivo. In addition, we addressed the questions (i) whether the expression of LTP reveals any age-dependence during the first two postnatal weeks, (ii) if the LTP is restricted to the stimulated barrel-related column and (iii) which cortical layers reveal LTP. To address these questions, we performed multi-channel extracellular electrophysiological recordings from barrels and barrel-related columns that were functionally identified by voltage-sensitive dye responses following single whisker stimulation.