Free SKILL.md scraped from GitHub. Clone the repo or copy the file directly into your Claude Code skills directory.
npx versuz@latest install hiyenwong-ai-collection-collection-skills-computational-taste-perceptiongit clone https://github.com/hiyenwong/ai_collection.gitcp ai_collection/SKILL.MD ~/.claude/skills/hiyenwong-ai-collection-collection-skills-computational-taste-perception/SKILL.md---
name: computational-taste-perception
description: '计算味觉感知方法论。混合神经元模型结合Hodgkin-Huxley生物物理保真度和Izhikevich脉冲神经元计算效率,模拟味觉传导通路。适用于味觉受体建模、脉冲神经网络、多尺度学习。触发词:味觉感知、脉冲神经网络、Hodgkin-Huxley、Izhikevich、味觉受体、STDP、taste perception、spiking neural network、gustatory。'
user-invocable: true
---
# Computational Taste Perception - 计算味觉感知
## 核心思想
混合神经元模型结合 Hodgkin-Huxley 生物物理保真度和 Izhikevich 脉冲神经元计算效率,实现味觉传导通路的精确模拟。
**来源:** arXiv:2510.00010
**效用:** 1.0
---
## 方法论
### 1. 混合神经元模型
**两层架构:**
| 层级 | 模型 | 用途 |
|------|------|------|
| 受体细胞 | Hodgkin-Huxley | 生物物理保真度 |
| 网络层 | Izhikevich | 计算效率 |
### 2. 味觉受体建模
**模态特异性受体动力学:**
- **T1R/T2R** - 甜/苦味受体
- **ENaC** - 咸味通道
- **PKD** - 酸味通道
**GHK 驱动离子电流:**
```python
def ghk_current(V, ion_conc_in, ion_conc_out, permeability, z=1):
"""
Goldman-Hodgkin-Katz 电流方程
Parameters:
-----------
V : float
膜电位
ion_conc_in/out : float
膜内外离子浓度
permeability : float
离子通透性
z : int
离子价态
Returns:
--------
I : float
离子电流
"""
F = 96485 # 法拉第常数
R = 8.314 # 气体常数
T = 310 # 温度 (K)
# GHK 方程
E = (R * T / (z * F)) * np.log(ion_conc_out / ion_conc_in)
I = permeability * z**2 * (F**2 * V / (R * T)) * \
(ion_conc_in - ion_conc_out * np.exp(-z * F * V / (R * T))) / \
(1 - np.exp(-z * F * V / (R * T)))
return I
```
### 3. 突触建模
**谷氨酸释放动力学:**
- Alpha 函数配置文件
- AMPA 受体贩运
- 磷酸化调节
**STDP 学习规则:**
```python
def stdp_update(w, delta_t, A_plus=0.1, A_minus=0.1, tau_plus=20, tau_minus=20):
"""
Spike-Timing-Dependent Plasticity
Parameters:
-----------
w : float
当前突触权重
delta_t : float
突触前-后脉冲时间差 (ms)
"""
if delta_t > 0:
# LTP: 突触前先于突触后
dw = A_plus * np.exp(-delta_t / tau_plus)
else:
# LTD: 突触后先于突触前
dw = -A_minus * np.exp(delta_t / tau_minus)
return w + dw
```
### 4. 多尺度学习
**两个维度:**
1. **时间脉冲同步** - van Rossum 指标
2. **组合群体编码** - 秩序模式
**van Rossum 距离:**
```python
def van_rossum_distance(spike_train1, spike_train2, tau=10):
"""
计算两个脉冲序列的van Rossum距离
"""
# 指数核滤波
def filter_spikes(spikes, tau):
t = np.linspace(0, 1000, 10000)
filtered = np.zeros_like(t)
for spike in spikes:
filtered += np.exp(-(t - spike) / tau) * (t >= spike)
return filtered
f1 = filter_spikes(spike_train1, tau)
f2 = filter_spikes(spike_train2, tau)
distance = np.sqrt(np.trapz((f1 - f2)**2))
return distance
```
---
## 实现框架
```python
import numpy as np
class HybridTasteNetwork:
"""混合味觉网络模型"""
def __init__(self, n_receptors=50, n_neurons=200):
self.n_receptors = n_receptors
self.n_neurons = n_neurons
# HH 受体模型参数
self.receptor_params = {
'T1R': {'g_Na': 120, 'g_K': 36, 'g_L': 0.3},
'T2R': {'g_Na': 100, 'g_K': 40, 'g_L': 0.3},
'ENaC': {'g_Na': 150, 'g_K': 30, 'g_L': 0.3},
'PKD': {'g_Na': 80, 'g_K': 35, 'g_L': 0.3}
}
# Izhikevich 网络参数
self.neuron_params = {
'a': 0.02, 'b': 0.2, 'c': -65, 'd': 8
}
def simulate_receptor(self, stimulus, receptor_type, dt=0.01):
"""HH 受体模拟"""
params = self.receptor_params[receptor_type]
# 简化的 HH 动力学
V = -65 # 静息电位
spikes = []
for i, I in enumerate(stimulus):
# HH 方程 (简化)
dV = (params['g_Na'] * (V - 50) + params['g_K'] * (-77 - V) +
params['g_L'] * (-54.4 - V) + I) * dt
V += dV
if V > 0: # 阈值
spikes.append(i * dt)
V = params['c']
return spikes
def simulate_network(self, receptor_spikes, duration=1000):
"""Izhikevich 网络模拟"""
a, b, c, d = self.neuron_params.values()
v = np.ones(self.n_neurons) * c
u = np.ones(self.n_neurons) * b * c
# 突触权重矩阵
W = np.random.rand(self.n_neurons, self.n_neurons) * 0.1
network_spikes = [[] for _ in range(self.n_neurons)]
for t in range(int(duration)):
# 输入电流
I = np.zeros(self.n_neurons)
# 从受体接收输入
for rs in receptor_spikes:
if t in rs:
I += np.random.rand(self.n_neurons) * 10
# Izhikevich 动力学
fired = v >= 30
v[fired] = c
u[fired] += d
dv = 0.04 * v**2 + 5 * v + 140 - u + I
du = a * (b * v - u)
v += dv
u += du
for n in np.where(fired)[0]:
network_spikes[n].append(t)
return network_spikes
```
---
## 应用场景
### 1. 味觉研究
- 味觉传导机制建模
- 受体动力学分析
- 味觉编码研究
### 2. 神经形态计算
- 混合神经元网络
- 多尺度学习
- STDP 实现
### 3. 仿生传感
- 电子舌设计
- 味觉传感器
- 食品质量检测
---
## 关键参数
| 参数 | 推荐值 | 说明 |
|------|--------|------|
| HH 时间步长 | 0.01 ms | 生物物理精度 |
| Izhikevich 时间步长 | 1 ms | 计算效率 |
| STDP tau | 20 ms | 可塑性窗口 |
| van Rossum tau | 10 ms | 脉冲距离 |
---
## 预期结果
1. **受体响应:** 模态特异性味觉受体脉冲
2. **网络同步:** 多尺度学习模式
3. **计算效率:** Izhikevich 比 HH 快 100x+
---
## 注意事项
- HH 模型计算成本高,用于受体层
- Izhikevich 模型高效,用于大规模网络
- STDP 参数影响学习收敛
- 需要验证生物保真度
---
## 参考文献
- arXiv:2510.00010 - Computational Advances in Taste Perception: From Ion Channels to Neural Coding
## Activation Keywords
- computational-taste-perception
- computational-taste-perception 技能
- computational-taste-perception skill
## Tools Used
- `read` - Read documentation and references
- `web_search` - Search for related information
- `web_fetch` - Fetch paper or documentation
## Instructions for Agents
Follow these steps when applying this skill:
### Step 1: 时间脉冲同步
### Step 2: 组合群体编码
### Step 3: 受体响应:
### Step 4: 网络同步:
### Step 5: 计算效率:
## Examples
### Example 1: Basic Application
**User:** I need to apply Computational Taste Perception - 计算味觉感知 to my analysis.
**Agent:** I'll help you apply computational-taste-perception. First, let me understand your specific use case...
**Context:** Apply the methodology
### Example 2: Advanced Scenario
**User:** Complex analysis scenario
**Agent:** Based on the methodology, I'll guide you through the advanced application...
### Example 2: Advanced Application
**User:** What are the key considerations for computational-taste-perception?
**Agent:** Let me search for the latest research and best practices...