Convolutional neural network assistance significantly improves dermatologists' diagnosis of cutaneous tumours using clinical images.

BACKGROUND: Convolutional neural networks (CNNs) have demonstrated expert-level performance in cutaneous tumour classification using clinical images, but most previous studies have focused on dermatologist-versus-CNN comparisons rather than their combination. The objective of our study was to evaluate the potential impact of CNN assistance on dermatologists for clinical image interpretation. METHODS: A multi-class CNN was trained and validated using a dataset of 25,773 clinical images comprising 10 categories of cutaneous tumours. The CNN's performance was tested on an independent dataset of 2107 images. A total of 400 images (40 per category) were randomly selected from the test dataset. A fully crossed, self-control, multi-reader multi-case (MRMC) study was conducted to compare the performance of 18 board-certified dermatologists (experience: 13/18 ≤ 10 years; 5/18＞10 years) in interpreting the 400 clinical images with or without CNN assistance. RESULTS: The CNN achieved an overall accuracy of 78.45% and kappa of 0.73 in the classification of 10 types of cutaneous tumours on 2107 images. CNN-assisted dermatologists achieved a higher accuracy (76.60% vs. 62.78%, P < 0.001) and kappa (0.74 vs. 0.59, P < 0.001) than unassisted dermatologists in interpreting the 400 clinical images. Dermatologists with less experience benefited more from CNN assistance. At the binary classification level (malignant or benign), the sensitivity (89.56% vs. 83.21%, P < 0.001) and specificity (87.90% vs. 80.92%, P < 0.001) of dermatologists with CNN assistance were also significantly improved than those without. CONCLUSIONS: CNN assistance improved dermatologist accuracy in interpreting cutaneous tumours and could further boost the acceptance of this new technique.

Repetitive microstimulation in rat primary somatosensory cortex (SI) strengthens the connection between homotopic sites in the opposite SI and leads to expression of previously ineffective input from the ipsilateral forelimb.

The primary somatosensory cortex (SI) receives input from the contralateral forelimb and projects to homotopic sites in the opposite SI. Since homotopic sites in SI are linked by a callosal pathway, we proposed that repetitive intracortical microstimulation (ICMSr) of neurons in layer V of SI forelimb cortex would increase spike firing in the opposite SI cortex thereby strengthening the callosal pathway sufficiently to allow normally ineffective stimuli from the ipsilateral forelimb to excite cells in the ipsilateral SI. The forelimb representation in SI in one hemisphere was mapped using mechanical and electrical stimulation of the contralateral forelimb, a homotopic site was similarly identified in the opposite SI, the presence of ipsilateral peripheral input was tested in both homotopic sites, and ICMS was used to establish an interhemispheric connection between the two homotopic recording sites. The major findings are: (1) each homotopic forelimb site in SI initially received short latency input only from the contralateral forelimb; (2) homotopic sites in layer V in each SI were interconnected by a callosal pathway; (3) ICMSr delivered to layer V of the homotopic SI in one hemisphere generally increased evoked response spike firing in layer V in the opposite homotopic site; (4) increased spike firing was often followed by the expression of a longer latency normally ineffective input from the ipsilateral forelimb; (5) these longer latency ipsilateral responses are consistent with a delay time sufficient to account for travel across the callosal pathway; (6) increased spike firing and the resulting ipsilateral peripheral input were also corroborated using in-vivo intracellular recording; and (7) inactivation of the stimulating site in SI by lidocaine injection or local surface cooling abolished the ipsilateral response, suggesting that the ipsilateral response was very likely relayed across the callosal pathway. These results suggest that repetitive microstimulation can do more than expand receptive fields in the territory adjacent to the stimulating electrode but in addition can also alter receptive fields in homotopic sites in the opposite SI to bring about the expression of previously ineffective input from the ipsilateral forelimb.

Differential Pattern of Interhemispheric Connections Between Homotopic Layer V Regions in the Forelimb Representation in Rat Barrel Field Cortex.

Layer V neurons in forelimb and shoulder representations in rat first somatosensory cortex (SI) project to the contralateral SI. However, few studies have addressed whether projections from specific subregions of the forelimb representation, namely forepaw, wrist, or forearm, terminate at homotopic sites in the contralateral SI. Neuroanatomical retrograde (cholera toxin B subunit [CT-B]) or anterograde (biodextran amine [BDA]) tracers were injected into physiologically identified sites in layer V in specific forelimb and/or shoulder representations in SI to examine the projection to contralateral SI in young adult rats (N = 17). Injection and target sites were flattened and cut in a tangential plane to relate labeling to the body map or cut along a coronal plane to relate labeling to cortical layers. Results indicate that layer V neurons project to cortical laminae II-VI in contralateral SI, with the densest labeling in layer V followed by layer III. In contrast, layer V neurons send sparse projections to layer IV. Furthermore, layer V neurons in wrist, forearm, and shoulder project to homotopic sites in contralateral layer V, while neurons in the forepaw representation project largely to sites in perigranular and dysgranular cortex adjacent to their homotopic territory. Our results provide evidence for a differential pattern of interhemispheric projections from forelimb and shoulder representations to the opposite SI and a detailed description of areal and laminar projection patterns of layer V neurons in the SI forelimb and shoulder cortices.

Forelimb amputation-induced reorganization in the ventral posterior lateral nucleus (VPL) provides a substrate for large-scale cortical reorganization in rat forepaw barrel subfield (FBS).

In this study, we examined the role of the ventral posterior lateral nucleus (VPL) as a possible substrate for large-scale cortical reorganization in the forepaw barrel subfield (FBS) of primary somatosensory cortex (SI) that follows forelimb amputation. Previously, we reported that, 6 weeks after forelimb amputation in young adult rats, new input from the shoulder becomes expressed throughout the FBS that quite likely has a subcortical origin. Subsequent examination of the cuneate nucleus (CN) 1 to 30 weeks following forelimb amputation showed that CN played an insignificant role in cortical reorganization and led to the present investigation of VPL. As a first step, we used electrophysiological recordings in forelimb intact adult rats (n=8) to map the body representation in VPL with particular emphasis on the forepaw and shoulder representations and showed that VPL was somatotopically organized. We next used stimulation and recording techniques in forelimb intact rats (n=5) and examined the pattern of projection (a) from the forelimb and shoulder to SI, (b) from the forepaw and shoulder to VPL, and (c) from sites in the forepaw and shoulder representation in VPL to forelimb and shoulder sites in SI. The results showed that the projections were narrowly focused and homotopic. Electrophysiological recordings were then used to map the former forepaw representation in forelimb amputated young adult rats (n=5) at 7 to 24 weeks after amputation. At each time period, new input from the shoulder was observed in the deafferented forepaw region in VPL. To determine whether the new shoulder input in the deafferented forepaw VPL projected to a new shoulder site in the deafferented FBS, we examined the thalamocortical pathway in 2 forelimb-amputated rats. Stimulation of a new shoulder site in deafferented FBS antidromically-activated a cell in the former forepaw territory in VPL; however, similar stimulation from a site in the original shoulder representation, outside the deafferented region, in SI did not activate cells in the former forepaw VPL. These results suggest that the new shoulder input in deafferented FBS is relayed from cells in the former forepaw region in VPL. In the last step, we used anatomical tracing and stimulation and recording techniques in forelimb intact rats (n=9) to examine the cuneothalamic pathway from shoulder and forepaw receptive field zones in CN to determine whether projections from the shoulder zone might provide a possible source of shoulder input to forepaw VPL. Injection of biotinylated dextran amine (BDA) into physiologically identified shoulder responsive sites in CN densely labeled axon terminals in the shoulder representation in VPL, but also gave off small collateral branches into forepaw VPL. In addition, microstimulation delivered to forepaw VPL antidromically-activated cells in shoulder receptive field sites in CN. These results suggest that forepaw VPL also receives input from shoulder receptive sites in CN that are latent or subthreshold in forelimb intact rats. However, we speculate that following amputation these latent shoulder inputs become expressed, possibly as a down-regulation of GABA inhibition from the reticular nucleus (RTN). These results, taken together, suggest that VPL provides a substrate for large-scale cortical reorganization that follows forelimb amputation.

Forelimb amputation-induced reorganization in the cuneate nucleus (CN) is not reflected in large-scale reorganization in rat forepaw barrel subfield cortex (FBS).

We examined reorganization in cuneate nucleus (CN) in juvenile rat following forelimb amputation (n=34) and in intact controls (n=5) to determine whether CN forms a substrate for large-scale reorganization in forepaw barrel subfield (FBS) cortex. New input from the shoulder first appears in the FBS 4 weeks after amputation, and by 6 weeks, the new shoulder input comes to occupy most of the FBS. Electrophysiological recording was used to map CN in controls and in forelimb amputees during the first 12 weeks following deafferentation and at 26 and 30 weeks post-amputation. Mapping was confined to a location 300 µm anterior to the obex where a medial-to-lateral row of electrode penetrations traversed through a complete complement of cytochrome-oxidase stained clusters (called barrelettes) that are associated with the representation of the glabrous forepaw digits and pads and adjacent non-cluster zones that are associated with the representation of the wrist, arm, and shoulder. Following amputation, non-cluster zones became occupied with new input from the body/chest and head/neck, while the cluster zone remained largely devoid of new input except at the border. A regression analysis comparing controls and amputees over the first 12 weeks post-amputation found significant differences for the total area of new input from the body/chest and head/neck in the non-cluster zones, while no significant differences were found for any new input into the cluster zone. When the averaged areas of a body-part representation were re-examined as a percentage of the averaged zonal area, a non-significant increase in new input from the body was observed within the cluster zone during post-amputation weeks 2-3 that returned to baseline in the subsequent weeks. In contrast, significant differences in averaged area of body-part representations for body/chest and head/neck were found in non-cluster zones over the first 12 weeks post-amputation. The present findings suggest that reorganization occurs only within the non-cluster zones whereby new input from the body/chest and head/neck moves in and occupies the deafferented territory immediately after amputation. Additionally, the lack of significant differences in new shoulder input in either cluster or non-cluster zones over the first 12 weeks after amputation suggests that CN provides an unlikely substrate for large-scale reorganization in the FBS.

Functional and structural organization of the forelimb representation in cuneate nucleus in rat.

We examined the physiological representation of the forelimb in the cuneate nucleus (CN) of forelimb-intact young adult rats (n=38) as the first part in a series of studies aimed at understanding the possible role that CN plays in delayed cortical reorganization that follows forelimb amputation. Metabolic labeling with cytochrome oxidase (CO) and electrophysiological mapping were used to examine the relationship between the structural and functional organization of CN. CN is a cylinder-shaped structure that lies bilaterally in the brainstem and extends nearly 4mm in the rostrocaudal direction. The forelimb is represented along the rostrocaudal extent. CN contains three zones; the rostral and caudal zones receive input largely from deep muscle and joint receptors and a middle zone, in the vicinity of the obex, receives input primarily from cutaneous receptors in the skin. The middle zone is somatotopically organized with the glabrous digits represented centrally, bordered on the medial side by ulnar wrist, ulnar forearm, and posterior upper arm representations; on the lateral side by radial wrist, radial forearm, and anterior upper arm representations; and on dorsal side by the dorsal digits and dorsal hand. The middle zone also contains well-defined CO-filled glomerular structures, called barrelettes, which are located within a homogenously stained field. The barrelettes are associated with the representation of the glabrous digits, with D5 represented most dorsal followed sequentially in a ventral-to-lateral direction by the representation of D4, D3, D2, and D1. The digit representations are topographically organized with the distal digit surface represented laterally with respect to the more medially lying proximal digit surface. The digit and palmar pads are also represented by barrelettes located on the medial side of CN. In contrast, the dorsal digit surfaces are represented dorsally and the dorsal hand is represented directly beneath the cuneate fasciculus, in a region devoid of barrelettes. The representations of the ulnar and radial wrist, forearm, and upper arm also lie within the homogeneously stained field in CN. The forelimb representation is bordered on the medial side by representation of trunk and hindlimb, and on the lateral side by representation of shoulder, ear, and head. While the present findings support and extend previous electrophysiological and anatomical studies of CN in the rat, they also provide a detailed physiological description of the functional organization of CN that is necessary for subsequent understanding of the functional reorganization of CN that may result following forelimb amputation.

Long-term effects of prenatal alcohol exposure on the size of the whisker representation in juvenile and adult rat barrel cortex.

Children of mothers who abused alcohol during pregnancy are often reported to suffer from growth retardation and central nervous system (CNS) abnormalities. The use of prenatal alcohol exposed (PAE) animal models has revealed reductions in body and brain weights as well as regional specific brain deficits in neonatal pups. Recently, we and others reported reductions in the size of the posteromedial barrel subfield (PMBSF) in first somatosensory cortex (SI) associated with the representation of the large mystacial vibrissae in neonatal rats and mice that were exposed to alcohol at various times during gestation. While these reductions in barrel field size were reported in neonates, it was unclear whether similar reductions persisted later in life or whether some catch-up might take place in older animals. In the present study, we examined the effect of PAE on measures of barrel field size in juvenile (6 weeks of age) and adult (7 months of age) rats; body and brain weights were also measured. Pregnant rats (Sprague-Dawley) were intragastrically gavaged during gestational days 1-20 with alcohol (6 g/kg) to simulate a binge-like pattern of alcohol consumption (Alc); 6 g/kg alcohol produced blood alcohol levels ranging between 207.4 and 478.6 mg/dl. Chow-fed (CF), pair-fed (PF), and cross-foster (XF) groups served as normal, nutritional/stress, and maternal controls, respectively, for juvenile rats; an XF group was not included for adult rats. The major findings in the present study are (i) PAE significantly reduced the size of the total barrel field in Alc juvenile rats (13%) and adult rats (9%) compared to CF controls, (ii) PAE significantly reduced the total averaged sizes of individual PMBSF barrels in juvenile (14%) and adult (13%) rats, (iii) PAE did not significantly alter the septal area between barrels or the barrel pattern, (iv) PAE significantly reduced body weight of juvenile rats but only in comparison to PF controls (18%), (v) PAE significantly reduced whole brain (8%) and forebrain (7%) weights of juvenile rats but not adult rats, (vi) no differences were observed in forebrain/PMBSF body ratios nor was forebrain weight correlated with PMBSF area, and (vii) PAE resulted in a greater reduction in anterior barrels compared to posterior barrels. These results suggest that the effects of PAE previously reported in neonate PMBSF areas persist into adulthood.

Early postnatal alcohol exposure reduced the size of vibrissal barrel field in rat somatosensory cortex (SI) but did not disrupt barrel field organization.

Prenatal alcohol exposure (PAE) has been shown to alter the somatosensory cortex in both human and animal studies. In rodents, PAE reduced the size, but not the pattern of the posteromedial barrel subfield (PMBSF) associated with the representation of the whiskers, in newborn, juvenile, and adult rats. However, the PMBSF is not present at birth, but rather first appears in the middle of the first postnatal week during the brain-growth spurt period. These findings raise questions whether early postnatal alcohol exposure might disrupt both barrel field pattern and size, questions that were investigated in the present study. Newborn Sprague-Dawley rats were assigned into alcohol (Alc), nutritional gastric control (GC), and suckle control (SC) groups on postnatal day 4 (P4). Rat pups in Alc and GC were artificially fed with alcohol and maltose-dextrin dissolved in milk, respectively, via an implant gastrostomy tube, from P4 to P9. Pups in the Alc group received alcohol (6.0 g/kg) in milk, while the GC controls received isocaloric equivalent maltose-dextrin dissolved in milk. Pups in the SC group remained with their mothers and breast fed throughout the experimental period. On P10, pups in each group were weighed, sacrificed, and their brains removed and weighed. Cortical hemispheres were separated, weighed, flattened, sectioned tangentially, stained with cytochrome oxidase, and PMBSF measured. The sizes of barrels and the interbarrel septal region within PMBSF, as well as body and brain weights were compared between the three groups. The sizes of PMSBF barrel and septal areas were significantly smaller (P<.01) in Alc group compared to controls, while the PMBSF barrel pattern remained unaltered. Body, whole-brain, forebrain, and hemisphere weights were significantly reduced (P<.01) in Alc pups compared to control groups. GC and SC groups did not differ significantly in all dependent variables, except body weight at P9 and P10 (P<.01). These results suggest that postnatal alcohol exposure, like prenatal exposure, significantly influenced the size of the barrel field, but not barrel field pattern formation, indicating that barrel field pattern formation consolidated prior to P4. These results are important for understanding sensorimotor deficits reported in children suffering from fetal alcohol spectrum disorder (FASD).

Altered keratin 17 peptide ligands inhibit in vitro proliferation of keratinocytes and T cells isolated from patients with psoriasis.

BACKGROUND: Identification of critical autoantigenic T-cell epitopes is key to developing antigen-based therapies for autoimmune diseases, including psoriasis. Our previous work demonstrated that 3 peptides on keratin 17 are able to stimulate peripheral blood lymphocytes of HLA-DRB1*07-positive patients with psoriasis and to serve as immunodominant T-cell epitopes. OBJECTIVE: We sought to determine antagonistic altered peptide ligands to psoriatic T cells with a down-modulatory effect in inhibiting keratinocyte proliferation. METHODS: Psoriatic altered peptide ligands were generated by single alanine residue substitutions at a critical T-cell receptor contact residue position. Antagonistic altered peptide ligands were identified by suppression screening of psoriatic T-cell activation and keratinocyte proliferation. RESULTS: Altered peptide ligands 119R and 355L can inhibit psoriatic T-cell activation more effectively than other altered peptide ligands, especially 355L, with inhibition of T-cell proliferation and the secretion of interferon gamma and interleukin 2 in parallel with the up-regulation of interleukins 4 and 10 as well as transforming growth factor-beta. In coincubation assay, altered peptide ligands 119R and 355L can down-regulate the function of psoriatic T cells more effectively than wild-type epitopes solely, but less effectively than altered peptide ligands solely. In prepulse assay altered peptide ligand 119R can down-regulate the activation of psoriatic T cells more effectively than in coincubation but less effectively as compared with altered peptide ligand 119R only. Altered peptide ligand 355L was also shown to have a similar presentation. T-cell culture supernatants (1:100) from the concentrations (10 microg.mL(-1) and 100 microg.mL(-1) with 119R, 100 microg.mL(-1) with 355L) were more effective than the other ratios in inhibiting keratinocyte proliferation. LIMITATIONS: This study had a relatively small sample size (52 patients and 48 healthy controls). CONCLUSION: Our findings show that the altered peptide ligands 119R (VAALEEANTELEVKI) and 355L (ENRYCVQASQIQGLI) are capable of inhibiting proliferative responses of psoriatic T cells and keratinocyte proliferation in vitro, at least, with enhanced helper T cell type 2 polarization. Thus, to our knowledge, this article is the first report of the demonstration of therapeutic activity of altered peptide ligands derived from keratin 17.

Prenatal alcohol exposure delays the development of the cortical barrel field in neonatal rats.

In-utero alcohol exposure produces sensorimotor developmental abnormalities that often persist into adulthood. The rodent cortical barrel field associated with the representation of the body surface was used as our model system to examine the effect of prenatal alcohol exposure (PAE) on early somatosensory cortical development. In this study, pregnant female rats were intragastrically gavaged daily with high doses of alcohol (6 gm/kg body weight) throughout the first 20 days of pregnancy. Blood alcohol levels were measured in the pregnant dams on gestational days 13 (G13) and G20. The ethanol treated group (EtOH) was compared to the normal control chowfed (CF) group, nutritionally matched pairfed (PF) group, and cross-foster (XF) group. Cortical barrel development was examined in pups across all treatment groups from G25, corresponding to postnatal day 2 (P2), to G32 corresponding to P9. The EtOH and control group pups were weighed, anesthetized, and perfused. Brains were removed and weighed with, and without cerebellum and olfactory bulbs, and neocortex was removed and weighed. Cortices were then flattened, sectioned tangentially, and stained with a metabolic marker, cytochrome oxidase (CO) to reveal the barrel field. Progression of barrel development was distinguished into three categories: (a) absent, (b) cloudy barrel-like pattern, and (c) well-formed barrels with intervening septae. The major findings are: (1) PAE delayed barrel field development by one or more days, (2) the barrel field first appeared as a cloudy pattern that gave way on subsequent days to an adult-like pattern with clearly demarcated intervening septal regions, (3) the barrel field developed differentially in a lateral-to-medial gradient in both alcohol and control groups, (4) PAE delayed birth by one or more days in 53% of the pups, (5) regardless of whether pups were born on G23 (normal expected birth date for non-alcohol controls) or as in the case for the alcohol-delayed pups born as late as G27, the barrel field was never present at birth suggesting the importance of postnatal experience on barrel field development, and (6) PAE did not disrupt the normal barrel field pattern, although both total body and brain weights were compromised. These findings suggest that PAE delays the development of the somatosensory cortex (SI); such delays may interfere with timing and formation of cortical circuits. It is unknown whether other nuclei along the somatosensory pathway undergo similar delays in development or if PAE selectively disrupts cortical circuitry.

Prenatal alcohol exposure (PAE) reduces the size of the forepaw representation in forepaw barrel subfield (FBS) cortex in neonatal rats: relationship between periphery and central representation.

Prenatal alcohol exposure (PAE) alters limb development that may lead to structural and functional abnormalities of the limb reported in children diagnosed with Fetal Alcohol Spectrum Disorder. To determine whether PAE alters the central representation of the forelimb we used the rodent barrel cortex as our model system where it was possible to visualize and quantitatively measure the size of the forepaw representation in the forepaw barrel subfield (FBS) in first somatosensory cortex. In the present study, we examined the effects of PAE on pattern and size of the forepaw and forepaw representation in FBS in neonatal rats at gestational day 32 that corresponds to postnatal day 9. Pregnant Sprague-Dawley rats were chronically intubated with binge doses of ethanol (6 g/kg) from gestational day 1 through gestational day 20. The offspring of the ethanol treated dams comprised the ethanol (EtOH) group. The effect of PAE on the EtOH group was compared with a nutritional-controlled pairfed (PF) group and a normal chowfed (CF) group. The ventral (glabrous) surface area of the forepaw digits, length of digit 2 through digit 5, and the corresponding glabrous forepaw digit representations in the FBS were measured and compared between treatment groups. In rats exposed to in utero alcohol, the sizes of the overall glabrous forepaw and forepaw digits were significantly reduced in EtOH pups compared to CF and PF pups; overall glabrous forepaw area was 11% smaller than CF controls. Glabrous digit lengths were also smaller in EtOH rats compared to CF controls and significantly smaller in digit 2 through digit 4. The glabrous digit representation in FBS was 18% smaller in the EtOH group when compared to the CF treatment. However, PAE did not produce malformations in the forepaw or alter the pattern of the forepaw representation in FBS; instead, PAE significantly reduced both body and brain weights compared to controls. Unexpectedly, little or no correlation was observed between the size of the glabrous forepaw compared to the size of the glabrous forepaw representation in the FBS for any of the treatment groups. The present findings of PAE-related alterations in sensory periphery and the central cortical representation may underlie deficits in sensorimotor integration reported among children with Fetal Alcohol Spectrum Disorder.

Genetic analysis of barrel field size in the first somatosensory area (SI) in inbred and recombinant inbred strains of mice.

We measured the combined area of posterior medial barrel subfield (PMBSF) and anterior lateral barrel subfield (ALBSF) areas in four common inbred strains (C3H/HeJ, A /J, C57BL /6J, DBA/2J), B6D2F1, and ten recombinant inbred (RI) strains generated from C57BL/6J and DBA/2J progenitors (BXD) as an initial attempt to examine the genetic influences underlying natural variation in barrel field size in adult mice. These two subfields are associated with the representation of the whisker pad and sinus hairs on the contralateral face. Using cytochrome oxidase labeling to visualize the barrel field, we measured the size of the combined subfields in each mouse strain. We also measured body weight and brain weight in each strain. We report that DBA/2J mice have a larger combined PMBSF/ALBSF area (6.15 +/- 0.10 mm(2), n = 7) than C57BL /6J (5.48 +/- 0.13 mm(2), n = 10), C3H/HeJ (5.37 +/- 0.16 mm(2), n = 10), and A/J mice (5.04 +/- 0.09 mm(2), n = 15), despite the fact that DBA/2J mice have smaller average brain and body sizes. This finding may reflect dissociation between systems that control brain size with those that regulate barrel field area. In addition, BXD strains (average n = 4) and parental strains showed considerable and continuous variation in PMBSF/ALBSF area, suggesting that this trait is polygenic. Furthermore, brain, body, and cortex weights have heritable differences between inbred strains and among BXD strains. PMBSF/ALBSF pattern appears similar among inbred and BXD strains, suggesting that somatosensory patterning reflects a common plan of organization. This data is an important first step in the quantitative genetic analysis of the parcellation of neocortex into diverse cytoarchitectonic zones that vary widely within and between species, and in identifying the genetic factors underlying barrel field size using quantitative trait locus (QTL) analyses.

Prenatal alcohol exposure alters the size, but not the pattern, of the whisker representation in neonatal rat barrel cortex.

Maternal alcohol exposure results in a variety of neurodevelopmental abnormalities that include cognitive and sensorimotor dysfunctions that often persist into adulthood. Many reports of central nervous system disturbances associated within a clinical diagnosis of fetal alcohol syndrome point toward disturbances in central information processing. In this study, we used the rat barrel field cortex as a model system to examine the effects of prenatal alcohol exposure (PAE) on the organization and size of the large whisker representation in layer IV of the posteromedial barrel subfield (PMBSF) in somatosensory cortex. Pregnant rats (Sprague-Dawley) were intragastrically gavaged daily with alcohol doses (6 gm/kg body weight) from gestational day 1 to day 20 in a chronic binge pattern which produced blood alcohol levels ranging between 260 mg/dl and 324 mg/dl. Chow-fed (CF), pair-fed (PF), and cross-foster (XF) groups served as normal, nutritionally matched, and maternal controls, respectively, for the ethanol-exposed (EtOH) treatment group. All pups were examined on gestational day 32 corresponding approximately to postnatal day 9. EtOH and control group pups were weighed, anesthetized, and perfused. Brains were removed and weighed, with and without cerebellum and olfactory bulbs, and the neocortex was removed and weighed. Cortices were then flattened, sectioned tangentially, and stained with a metabolic marker-cytochrome oxidase-to reveal the barrel field. A subset of 27 cortical barrels, associated with the representation of the large whisker pad, was selected to examine in detail. The major results were: (i) the total barrel field area comprising the PMBSF was significantly reduced in EtOH (by 17%) and XF (by 16%) pups compared with CF pups, (ii) the sizes of individual barrels within the PMBSF were also significantly reduced in EtOH (16%) and XF (18%) pups, (iii) the septal region between barrels was also significantly reduced in EtOH (18%) and XF (12%) pups, (iv) anteriorly located barrels underwent greater reduction in size relative to the posteriorly located barrels, (v) body weights were also significantly reduced in EtOH (21%) and XF (27%) pups, (vi) total brain weight [with and without (forebrain) cerebellum/olfactory bulbs] and cortical weights were also significantly reduced in EtOH (total brain weight 15%, forebrain weight 16%, cortical weight 15%) and XF (18%, 19%, 20%) pups, and in contrast (vi) neither the overall barrel field pattern nor the pattern of individual barrels in the PMBSF was altered. These findings suggest that PAE reduces body and brain weight as well as the central cortical representation of the whisker pad, while leaving the overall barrel field pattern unperturbed. While these results might appear to support a miniaturization hypothesis (smaller PMBSF, smaller brain, smaller body weight), PAE also shows regional vulnerability within the PMBSF whereby anteriorly located barrels are most affected.

Delayed reorganization of the shoulder representation in forepaw barrel subfield (FBS) in first somatosensory cortex (SI) following forelimb deafferentation in adult rats.

We previously reported that 6-16 weeks after forelimb amputation in adult rats, neurons in layer IV of rat first somatosensory cortex (SI) in the forepaw barrel subfield (FBS) associated with the representation of the forepaw became responsive to new input from the shoulder (Pearson et al. 1999). These new shoulder-responsive sites in deafferented FBS had longer evoked response latencies than did sites in the shoulder representation located in the posterior part of the trunk subfield, hereafter referred to as the original shoulder representation. Furthermore, projection neurons in the original shoulder representation in both intact and deafferented adults did not extend their axons into the FBS, and ablation of the original shoulder representation cortex and/or the second somatosensory cortex (SII) failed to eliminate new shoulder input in the deafferented FBS (Pearson et al. 2001). These results led us to conclude that large-scale reorganization in FBS quite likely involved a subcortical substrate. In addition, the time course for large-scale cortical reorganization following forelimb amputation was unknown, and this information could shed light on potential mechanisms for large-scale cortical reorganization. In the present study, we extended our previous findings of large-scale cortical reorganization in the FBS by investigating the time course for reorganization following forelimb amputation. The major findings are: a) deafferented forelimb cortex remained unresponsive to shoulder stimulation during the 1st week following forelimb amputation; b) new responses to shoulder stimulation were first observed in deafferented forelimb cortex 2-3 weeks after forelimb amputation; however, the new shoulder input was restricted to locations in the former forearm cortex; c) islet(s) of new shoulder representation were first observed in deafferented FBS 4 weeks after amputation; these islets occupied a larger percentage of FBS in subsequent weeks; d) portions of FBS remained unresponsive as many as 4 months after deafferentation (maximum time examined between amputation and recording); and e) the increase in total size of the shoulder representation appeared to result from the establishment of new shoulder representations that were often discontinuous from the original shoulder representation. These findings provide evidence that forelimb amputation results in delayed reorganization of the FBS and we describe possible mechanisms and substrates underlying the reorganization.

Removal of GABAergic inhibition alters subthreshold input in neurons in forepaw barrel subfield (FBS) in rat first somatosensory cortex (SI) after digit stimulation.

Our objective was to test the hypothesis that suppression of GABAergic inhibition results in an enhancement of responses to stimulation of the surround receptive field. Neurons in the forepaw barrel subfield (FBS) in rat first somatosensory cortex (SI) receive short latency suprathreshold input from a principal location on the forepaw and longer latency subthreshold input from surrounding forepaw skin regions. Input from principal and surround receptive field sites was examined before, during, and after administration of the GABA(A) receptor blocker bicuculline methiodide (BMI) (in 165 mM NaCl at pH 3.3-3.5). In vivo extracellular recording was used to first identify the location of the glabrous forepaw digit representation within the FBS. In vivo intracellular recording and labeling techniques were then used to impale single FBS neurons in layer IV as well as neurons in layers III and V, determine the receptive field of the cell, and fill the cell with biocytin for subsequent morphological identification. The intracellular recording electrode was fastened with dental wax to a double-barrel pipette for BMI iontophoresis and current balance. A stimulating probe, placed on the glabrous forepaw skin surface, was used to identify principal and surround components of the receptive field. Once a cell was impaled and a stable recording was obtained, a stimulating probe was placed at a selected site within the surround receptive field. Single-pulse stimulation (1 Hz) was then delivered through the skin probe and the percentage of spikes occurring in 1-min intervals before BMI onset was used as a baseline measure. BMI was then iontophoresed while the periphery was simultaneously stimulated, and spike percentage measured during and after BMI ejection was compared with the pre-BMI baseline. The major findings are: (1) suppression of GABAergic inhibition enhanced evoked responses to firing level from sites in surround receptive fields in 65% of the cells ( n=17); (2) evoked responses were rapidly elevated (within 1 min) to suprathreshold firing in the presence of BMI in 31% of the cells; (3) GABAergic inhibition was reversible [suprathreshold spiking gradually reversed to subthreshold excitatory postsynaptic potentials (EPSPs) in 45% of the cells tested]; (4) BMI altered the stimulus-evoked and non-stimulus-evoked firing pattern in SI neurons from single spikes to burst patterns in all tested cells; and (5) iontophoresis of NaCl (165 mM) without BMI was ineffective in altering evoked responses in control cells ( n=4). The present findings support the notion that subthreshold input from surround receptive fields is one possible mechanism for rapid cortical reorganization in barrel cortex and that GABAergic inhibition may regulate its expression. Possible corticocortical and thalamocortical substrates for subthreshold input to reach barrel neurons are discussed.