The present study is the first to perform in vivo patch-clamp recordings from ACC neurons of adult mice under urethane anesthesia and to systematically characterize the action potential (AP) properties of layer II/III pyramidal neurons. From single labeled neuron morphological analyses, we confirmed recording positions. We chose to record from layer II/III neurons in the ACC because (1) our previous work mostly focused on synaptic transmission and plasticity of ACC layer II/III [18, 20, 21]; (2) most neurons located in layer II/III are pyramidal neurons; and (3) neurons in layer II/III receive robust sensory input. We found that three major electrophysiological classes of pyramidal neurons could be distinguished according to their firing patterns and the shapes of their APs: RS, IM and IB neurons. Finally, we applied cutaneous noxious pinch and innocuous brush stimuli to the ispilateral and/or contralateral hind paws whilst recording from each type of neuron. We revealed that the pinch-evoked responses in IB cells had significantly higher frequencies than those in RS and IM, although brush stimuli did not activate any cell type. In contrast, there were no significant differences between ispilateral and contralateral evoked-responses across all cell types.
Based on AP shape and specific discharge patterns, our current study recorded from three identified types of pyramidal neurons within the ACC, which composed 10, 62 and 28% of our recordings respectively. In contrast, using the same classification protocols, our previous in vitro spike study  demonstrated different population distributions for ACC RS, IM and IB pyramidal neurons in layer II/III of adult mice (25, 44 and 31%, respectively). In comparison with the in vitro study, IM neurons were more frequently observed (62% in vivo vs. 44% in vitro). On the other hand, the number of RS neurons observed under in vivo conditions (10%) was lower than under in vitro conditions (25%) although the percentage of IB (28%) neurons observed in our in vivo study was similar to those observed in vitro (31%). The population ratios of RS and IM cells under in vivo and in vitro conditions were significantly different (P < 0.01, χ2 test). One key conclusion from these studies is that in vitro brain slice preparations is useful for studying the spike properties of different types of ACC pyramidal cells in vivo.
It is important to note that some of the variations in the population ratios between in vivo and in vitro observations may be related to recording conditions including (i) anesthetic effects and (ii) variations in recording temperatures. The generation and shape of action potentials depend on sodium (Na), potassium (K) and calcium (Ca) channels [29, 30]. The main criteria distinguishing IM cells from RS cell is the presence of the afterdepolarization (ADP), which is generally composed of Na, T- and R-type Ca currents . Compared with in vitro conditions, in vivo states are more likely to maintain neurons in a healthy condition, and thus these channels may be more active than under in vitro conditions. In addition, it has been reported that urethane can affect various channels, including Na channels [31–33]. Although there were no differences in slow oscillation frequencies produced by urethane in the three types of cells (Fig. 2C), we cannot eliminate the effects of anesthesia on firing pattern classifications. Furthermore, our previous in vitro study  was conducted at room temperature, whereas the present study used recording temperatures ranging from 36-37°C. These differences in temperature may affect channel gating, which could explain a difference in ADP amplitude and hence the proportion of cells that are RS vs IM.
Spontaneous and evoked responses
We observed that the spontaneous APs frequencies in IB neurons were significantly higher than those in RS and IM neurons (Fig. 4A). It has been reported that the generation of AP bursting discharges are mainly dependent on Na and K channel activity [29, 30]. In fact, by raising extracellular K concentrations, or through potentiation of the persistent Na current with the Na channel toxin ATX II, a subset of cortical pyramidal neurons can be induced to switch from repetitive single spiking to a burst firing mode . Additionally, the blocking of M-current, which is a voltage-gated potassium current (current through Kv7 channels) and a low-threshold non-inactivating potassium current, neuronal firing can be switched from tonic discharge to an intermittent firing mode in the entorhinal cortex  and hippocampus . Therefore, it is possible that individual types of pyramidal neurons in the ACC may have different expressions of AP related channels.
We also found that IB neurons responded significantly more to noxious pinch stimuli (333%) than RS (185%) and IM cells (185%) (Fig. 6B). A possible explanation for this may be, in addition to differences in their intrinsic properties, that they exhibit differences in sensory input and anatomical features of individual types. It has been reported that IB and RS neurons in the neocortex have different characteristics in excitatory and inhibitory input [36–38]. IB neurons mainly receive excitatory input and rarely receive GABAergic inhibitory input . Furthermore, they receive smaller and slower rises and decays of GABAergic inputs than RS cells [36, 37]. Anatomically, IB neurons have larger cell bodies and longer, thicker apical dendrites than those in RS cells [36–38]. The functional implications of their intrinsic electrophysiological properties, their responses to various sensory stimuli and their morphological features need to be further investigated.
Neurons in layer II/III of the ACC receive sensory information from the medial thalamus, and in turn project to layer V neurons [7, 8, 39–41]. Layer V neurons in turn form synapses with various cortical areas including the amygdala, a structure critical for emotional fear and anxiety, and also the motor cortex, where they can generate motor responses such as emotional vocalizations or trigger aversive behaviors [4, 42, 43]. Taking into considerations of the different intrinsic physiological properties and pinch-evoked responses obtained in this study, it is possible that different classes of pyramidal neurons within the ACC contribute in different ways to the information processing that takes place in the ACC. In fact, when noxious stimuli was applied, RS and IM cells displayed phasic firing at the onset of the stimuli, whereas IB cells displayed a tonic firing pattern, acting like powerful amplifiers during stimulation (Fig. 5B). Additionally, noxious pinch stimuli induced increases in evoked AP frequency greater than 200% of control in 50% of IB neurons (Fig. 6A). Therefore, the generation of high-frequency burst discharges in IB neurons may strongly influence the responses of postsynaptic neurons and the operation of local cortical networks, and also may play important roles in various emotional and motor responses.